LOGIN | HOME | CONTACT

HEALTHCARE OPERATIONS RESEARCH FOCUS AREA

Examining the catchment area of a neurosurgery service at a tertiary care facility in the southeast and the potential value of a health information exchange (HIE)
This project entails quantifying the number of patients received as transfer patients to the Emory Neurosurgery Service over a one year period of time and identifying the top sending institutions. The hypothesis is that selectively establishing an HIE between the top sending institutions to Emory Neurosurgery could decrease costs to the nation’s health care system. The goal would be to extrapolate any findings in order to make a statement about the impact the role out of HIEs for transfer patients (selectively and thoughtfully between centers where there are high volumes of transfer patients) broadly across the nation might have on the nation’s ever growing healthcare costs. This argument would be made by making some assumptions about a few tests potentially ordered twice within a short window of time (i.e. at the sending institution and at Emory within an 8 or 12 hour window of time with a low suspicion for any significant change within that time interval) with regular frequency (perhaps an EKG on a neurosurgery ICU admission with no cardiac complaint, for example) on transfer patients and the cost of these tests (more specifically the cost associated with the professional fees for the interpretation of the tests as there is no real other incremental cost to a payor for an additional test done while a patient is admitted inpatient in the DRG based prospective payment system). We would then arrive at some quantification of the value of establishing an interface that would facilitate the exchange of test results in lieu of repeating the tests. The underpinning logic would be that an HIE could potentially reduce the frequency with which tests done at the sending facility are duplicated at the receiving facility and thereby lower costs. The amount by which costs could potentially be lowered would serve to estimate the potential value that could be created by establishing an HIE presuming an HIE could successfully change the propensity for a physician to repeat a study available and easily accessible from an outside hospital facility.

Return to the Top of the Page

Hospital throughput & patient flow analysis
There is a wealth of recent patient through put flow data with time stamps and a database of transfer patients from the transfer center (with data on acceptance, denial, no bed available, etc.) available to us. The information is available as Emory has started using eBeds, a McKesson product, which is designed to raise insight into hospital bed management operations (see below from McKesson marketing materials):

  • Provides more effective capacity management by automating bed management functions to increase patient throughput and revenue
  • Boost productivity by streamlining hospital bed management and transportation processes and workflow
  • Enhance patient and caregiver satisfaction by reducing patient wait times for admission and transfers and improving your caregivers’ ability to schedule and track patients
  • Improve hospital bed management with a centralized, enterprise-wide bed board that provides real-time bed tracking statistics and modeling

The high level concept is to put a price tag on the opportunity cost (inability to accommodate demand for beds by transfer patients that ultimately get go to another facility because a bed is not available) of inefficient management of patient transfers or said another way the cost of poor capacity utilization. We will also determine the value that can be created by improving delays in throughput by 10, 20, 30, 40 percent,...(i.e. sensitivity analysis), by addressing bottlenecks and other patient flow problems. Two of the initial focuses for this effort are detailed below:

  1. determine volume of potential incoming patients (transfers) to neurosurgery and evaluate how many patients are turned away (patient accepted but no bed available) secondary to capacity constraints and the potential cost of turning away these patients (lost DRG revenue for an incremental admission of a particular diagnosis)
  2. examine the flow of admitted patients to neurosurgery through their hospital course from admission to discharge. 3 potential inputs into the patient funnel...direct admissions, emergency department admissions, elective surgeries, transfers from outside hospitals. there is a staging area for non critical care patients called care initiation where some patients initially go. post operative patients stay in the post anesthesia care unit until they are recovered and there is a bed available for them. er patients to be admitted stay in the er until a bed is available in an appropriate care setting be it floor, icu, care initiation, etc. transfer patients cannot leave the outside hospital facility until a bed is available and held for the transfer patient. patients then move through whatever care settings are necessary for their course of care (i.e. icu to intermediate care to floor, or whatever combination might be required for a given patient...then discharge, if they do not die in the hospital, to a different care setting, like home, skilled nursing facility, rehab facility, long term care facility, or hospice). bottle necks could occur for any number of reasons as a patient moves through the hospital. one of the primary ones that could be analyzed would be when the patient has discharge orders, but cannot be moved out because there is no bed available to move them into. any time this happens there is potentially a bed that otherwise would be open that the hospital cannot take advantage of [i.e. by giving the go-ahead for a transfer patient to transport from an outside hospital] because they have no where to which they can move the patient currently in the room, that no longer require that level of care. it is possible that a patient may be waiting for radiology to do a post operative scan for example before they can be moved out of the icu, but I do not think the raw data would allow analysis at this level.


     Return to the Top of the Page