Patent Application
20150148617
The present disclosure generally relates to a method and system for providing an early indication or warning of impending patient deterioration while reducing nuisance alarms. More specifically, the present disclosure relates to a method and system that is able to select one of a plurality of early warning algorithms to most accurately provide an early indication or warning, of an impending patient deterioration.
Presently available patient monitoring systems are able to monitor a relatively large number of different physiological data parameters obtained from a patient. in addition to monitoring the physiological data obtained from the patient, the patient monitors are able to retrieve stored information about the patient from an electronic health record. Further, the patient monitor allows the user to enter diagnostic information about the patient into the monitor. Thus, a clinician viewing the patient monitor is presented with what can be an overwhelming amount of changing data.
Because automatic monitoring of patients is becoming increasingly prevalent, various challenges have been identified. One of the principal challenges that result from automatic patient monitoring is alert fatigue. Alert fatigue is referred to as the condition in which clinicians become desensitized to clinical alerts because of the high probability that the alerts are not of actual clinical significance. Although one way to address this problem is to raise alert thresholds, this reduces sensitivity and increases the likelihood of failing to detect patients in deterioration.
As a result of the large amount of data available at the patient monitor, different types of early warning, algorithms have been developed to analyze the various different types of data and provide a single indication of the current health status of the patient. These various different algorithms operate utilizing different numbers and types of information to generate the assessment of the patient's overall health. The use of such algorithms allows for an early warning of the deteriorating health of the patient based upon the multiple parameters being monitored.
Although various different types of early warning algorithms have been developed to process various combinations of patient physiological data, patient demographic data and patient diagnostic data, each of these algorithms is specifically tailored for a certain type of available data as well as the type of patient being monitored. As an illustrative example, early warning algorithms have been developed for neonatal patients, which differ from algorithms developed for elderly or middle-aged patients. Therefore, in order for a patient monitor to accurately assess the overall health of a patient by utilizing an early warning algorithm, the patient monitor must be operating utilizing an algorithm that is appropriate for the patient being monitored.
The present disclosure relates to a method and system for selecting and downloading an early warning algorithm for use with a patient monitor. The system and method identifies one of a plurality of an early warning algorithms that is most desirable for use with a patient monitor based upon at least one of patient demographic data, patient physiological data and patient diagnostic data that is available for the patient.
In accordance with the present disclosure, a patient monitor that is being used to monitor the status of a patient communicates with an algorithm selection device. The algorithm selection device accesses a computer database that includes the plurality of early warning algorithms. Based upon information available about the patient, the method computes a ranking of the early warning algorithms.
Once the plurality of early warning algorithms have been ranked, the method selects one of the early warning algorithms for use with the patient monitor. The selection of the early warning algorithm is based at least partially on one of the patient demographic data, the patient physiological data and the patient diagnostic data. The selection of the early warning algorithm can be done automatically by the algorithm selection device or can be done manually by a user.
After the most desirable early warning algorithm has been selected, the patient monitor operates using the selected early warning algorithm. In one embodiment, the selected early warning algorithm is downloaded to the patient monitor. Once the patient monitor receives the downloaded early warning algorithm, the patient monitor operates to monitor the patient and provide data to the early warning algorithm for analysis.
In another alternate embodiment, the selected early warning algorithm could also be executed in an algorithm server separate from the patient Monitor. In such a configuration, the patient monitor and the algorithm server would be configured to operate over the hospital network. Data obtained from the patient would be provided to the early warning algorithm over the hospital network for evaluation by the early warning algorithm residing on the algorithm server.
When the selection of the early warning algorithm is being made, the system and method utilizes various different parameters to determine which of the early warning algorithms may be most desirable for monitoring the patient. This selection can include determining What types of patient physiological data. In addition, the selection process can also be based on the types of patient demographic data and patient diagnostic data that may be available from an electronic health record for the patient. Once the early warning algorithms have been analyzed, the method displays the identified early warning algorithms to a user and allows the user to select one of the most preferred algorithms. The selection process can include ranking each of the early warning algorithms and displaying the ranking to a user such that the user can select the algorithm based upon the determined rankings.
The present disclosure also relates to a patient monitoring system for monitoring the status of a patient. The patient monitoring system includes at least one patient monitor that includes a display device and at least one sensor connected to the patient to obtain physiological data from the patient. The patient monitoring system further includes a computer database that includes a plurality of early warning algorithms. Each of the early warning algorithms operates to predict the deteriorating health of a patient based upon multiple parameters.
The patient monitoring system further includes an algorithm selection device that is in communication with the computer database and the patient monitor. The algorithm selection device includes a processor that is programmed to select one of the plurality of early warning algorithms based at least partially on the physiological data obtained from the patient. Further, the processor downloads the selected early warning algorithm to the patient monitor such that the patient monitor operates utilizing the downloaded early warning algorithm.
The drawings illustrate the best mode presently contemplated of carrying out the disclosure. In the drawings:
Each of the patient monitors 12 further includes a display 18 that allows a controller 22 contained within the patient monitor 12 to present information to a user. Although the display 18 is shown separate from the patient monitor 12, it should be understood that the display 18 could be incorporated within the patient monitor while operating within the scope of the present disclosure.
In addition, each patient monitor 12 further includes an input device 20 that allows the user to enter various different data into the patient monitor 12 as well as to make selection choices based upon information shown on the display 18. The input device 20 can be various different types of components, such as a keyboard, mouse or any other device that allows information to be entered into the patient monitor 12. Further, it should be understood that the display 18 and input device 20 could be combined into a single device, such as a touch screen.
The patient monitor 12 includes a controller 22 and a memory device 24. The memory device 24 can store operating alarm limits, operating parameters as well as an algorithm for generating early warning alarms to a user. The controller 22 carries out the algorithm stored within the memory device 24 to generate alarms to a user, either through the display 18 or through other types of warning devices.
As illustrated in
The algorithm selection device 28 is also in communication with the hospital information system 34 through the communication network 26. The hospital information system 34 includes a database of patient information that includes a significant number of electronic health records (EHR) 36. Each EHR is stored for each of the patients 14 being monitored, as well as for all other patients that have been monitored in the hospital. The EHR 36 includes all of the information obtained from the patient over the lifetime of treatment of the patient at the facility. The EHR can include patient demographic data, historic patient physiological data and patient diagnostic data. As an illustrative example, the patient demographic data can include information relating to the patient's age, gender, family health history, race, height, weight and other types of parameters that are typically stored within an EHR 36. The patient diagnostic data can include information relating to past diagnostic assessments made for the patient, such as information relating to whether the patient is diabetic, obese, septic, or any other diagnostic assessments that have been made for the patient and may affect the current monitoring of the patient.
In addition to being accessible by the algorithm selection device 28, the hospital information system 34 can also be accessed by each of the patient monitors 12 through the communication network 26. Each of the patient monitors 12 may utilize the patient information contained in the EHR 36 to enhance the monitoring of the patient.
As can be seen and understood in
As Shown in Table 1, the Rothman index (adult) is an early warning algorithm that is particularly useful with adult patients based upon episodic data. The Rothman index utilizes twenty-six different parameters in generating an overall index. The twenty-six patient indicators are all obtained from the EHR 36 for the patient. These parameters obtained from the EHR 36 can include historic vital sign information, pulse oximeter data, lab values, a Braden score, individual nursing assessments, as well as other data included in the EHR 36. Based upon the twenty-six parameters obtained, the Rothman early warning algorithm generates an index score that provides an early warning indication of the patient's health.
In addition to the Rothman index for adult patients, the database 30 further includes an early warning algorithm that generates a Rothman index for pediatric patients. The Rothman index for pediatric patients also uses twenty-six different parameters. However, since pediatric patients have different parameters that may indicate deteriorating health, it is desirable to use the early warning algorithm that generates the Rothman index for pediatric patients when the patient monitor is being used to monitor a pediatric patient.
The next early warning algorithm 32 shown in
The next algorithm shown in
The final algorithm in
As can be understood by Table 1 and the above description, the algorithm selection device 28 shown in
Set forth below in Table 2 is a sample weighting system that can be utilized by the algorithm selection device 28 in selecting which of the early warning algorithms 32 is most beneficial for monitoring an individual patient. In Table 2, a weight is assigned for each type of the patient demographic data and patient physiological data that is available from either the patient monitor 12 or the ERR 36. As can be seen below, the type of data, whether continuous or episodic, is a factor used to select between the early warning algorithms. Based upon a combined score of the parameters that are available for use, the algorithm selection device can use a compiled score for each algorithm to rank and ultimately select the one or more early warning algorithms that are most applicable for the patient.
Once the most desirable algorithm is selected, the algorithm utilized to control the operation of the patient monitor 12. In one embodiment, the selected early warning algorithm is downloaded to the memory device 24 contained within the patient monitor 12. Thus, the patient monitor 12 can be delivered without any pre-stored algorithm. Further, the patient monitor 12 can obtain additional or alternate early warning algorithms 32 when the patient monitor is used to monitor different types of patients within a hospital environment. In this manner, the patient monitor 12 can be configured during use and most effectively provide an early warning algorithm 32 based upon the patient being monitored.
In an alternate embodiment, the selected early warning algorithm can be operated on an algorithm server that is separate from the patient monitor. The algorithm server could be part of the algorithm selection device of could be a separate component. In this embodiment, the patient monitor communicates the monitored patient physiological data to the algorithm server over the hospital network. The early warning algorithm would operate using the physiological data and the other available patient diagnostic or demographic data and communicate back to the patient monitor through the hospital network.
Although the selection of the algorithm is made in the algorithm selection device 28, it is contemplated that such determination could also be made at the patient monitor 12 using the controller 22. However, the embodiment and configuration shown in
Once this data is obtained at the patient monitor, the existence of the physiological data is reported up to the algorithm selection device through the communication network, as indicated in step 52. Once the algorithm selection device receives information about the type of physiological data that is available through the sensors, the algorithm selection device communicates to the hospital information system and determines what type of diagnostic data and demographic data are available for the patient from the electronic health record, as indicated in step 54. As previously described, various different types of diagnostic data and demographic data can be available for the patient in the EHR. However, if the patient has not been in a clinical or hospital environment, the EHR may not include any information and thus various different early warning algorithms may not be applicable for use with that patient.
Once the algorithm selection device has obtained information about the types of physiological data, demographic data and diagnostic data that is available for the patient being monitored at the patient monitor, the algorithm selection device determines in step 56 which of the plurality of early warning algorithms may be relevant for the patient. If the EHR is complete and multiple sensors are being used with the patient, more than one of the early warning algorithms may be particularly useful in monitoring a patient. Alternatively, if the EHR is not complete and only very little physiological data is being obtained from the patient, only one of the multiple early warning algorithms may be desirable for use with the patient.
In step 58, the algorithm selection device 28 displays a list of early warning algorithms that could be used with the patient monitor. This list of possible early warning algorithms can include a ranking indicating which of the early warning algorithms appears to be most desirable for use with the patient monitor. This ranking can be determined based upon the scoring system shown in Table 2 or can be based upon other parameters. The list of possible early warning algorithms and a ranking of the algorithms can be presented to the user either on a display 61 associated with the algorithm selection device or on the display 18 associated with the patient monitor. Since the user most likely will be at the patient monitor 12, it is contemplated that the listing of possible early warning algorithms will be presented on the display 18 at the patient monitor.
Alternatively, instead of presenting information to the user, the algorithm selection device 28 could automatically select the most preferred early warning algorithm 32 without requiring any user input. Although the automatic selection of the early warning algorithm may remove any incorrect selection by the user, it is contemplated that presenting the selection information to the user would allow the user to make informed decisions about the type of algorithm being utilized to operate the patient monitor.
In step 60, the algorithm selection device receives a selection by the user of the desired early warning algorithm for use in the patient monitor. The selection made by the user can be through a user input device 20 associated with the patient monitor or through a similar input device associated with the algorithm selection device 28 shown in
Once the user selects the early warning algorithm 32, the early warning algorithm is downloaded to the patient monitor, as indicated in step 62. The early warning algorithm is downloaded through the communication network 26 and is stored within a memory device 24 associated with the patient monitor. Once the early warning algorithm has been downloaded, the patient monitor 12 operates utilizing the early warning algorithm and generates patient alerts/warnings based upon the operation of the early warning algorithm.
If the patient monitor is used with a different type of patient or with a different grouping of sensors, the patient monitor can again communicate back to the algorithm selection device for the algorithm selection device to select the early warning algorithm that is most relevant, as was described with reference to step 56 shown in
This written description uses examples to disclose the invention, including the best mode, and also to enable any person skilled in the art to make and use the invention. The patentable scope of the invention is defined by the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal languages of the claims.
