Forecasting macrophage activation syndromes

Abstract

Technologies are disclosed for determining or predicting the occurrence of a macrophage activation syndrome, such as hemophagocytic lymphohistiocytosis (HLH). A detection of the emergence of and/or a reliable estimation of the likelihood of future significant macrophage activation syndromes, such as HLH, may be determined or predicted from a time series of laboratory and physiologic values to be measured in a patient. Root mean square of successive deviations (RMSSD) is utilized as a surrogate non-parametric measure of the high-frequency power spectral density (PSD) to identify strong statistical associations with the presence and/or near-term future emergence of macrophage activation syndromes. Utilizing these input variables, a model having satisfactory predictive accuracy is constructed using linear discriminant analysis (LDA), gradient boosting, random forest (RF), neural network, logistic regression, or the like, and may be used for the prediction.

What is claimed is: 
     
         1 . A method comprising:
 receiving one or more measurements of a subject, the one or more measurements comprising:
 one or more first laboratory or physiologic measurements and one or more corresponding date-time coordinates; and 
 one or more second laboratory or physiologic measurements and a respective date-time coordinate, the one or more second laboratory or physiologic measurements being less recent than the one or more date-time coordinates corresponding to the one or more first laboratory or physiologic measurements; 
   retrieving a time series including at least the one or more first laboratory or physiologic measurements and the one or more second laboratory or physiologic measurements, the time series constructed based on the one or more corresponding date-time coordinates of the one or more first laboratory or physiologic measurements and the respective date-time coordinate of the one or more second laboratory or physiologic measurements;   determining a probability of a macrophage activation syndrome for the subject by determining, for the time series, high-frequency components of a power spectrum or a surrogate measure of a high-frequency band of the power spectrum, the probability determined by using one or more predictive models;   determining that the probability of the macrophage activation syndrome exceeds a threshold associated with a condition of the subject, with the one or more first laboratory or physiologic measurements, with the one or more second laboratory or physiologic measurements, or with other laboratory or physiologic measurements;   in response to determining the probability of the macrophage activation syndrome exceeds the threshold, identifying a recommendation for a treatment or a therapy for the subject; and   transmitting a notification to a caregiver indicating that the subject has the probability, corresponding to a future time interval, of the macrophage activation syndrome exceeding the threshold, the notification comprising the recommendation.   
     
     
         2 . The method of  claim 1 , wherein the surrogate measure of the high-frequency band of the power spectrum comprises a root mean square of successive deviations (RMS SD). 
     
     
         3 . The method of  claim 1 , wherein the threshold is pre-determined, specific to the subject, or established for issuing an alarm to the subject or the caregiver, and wherein the threshold comprises one or more thresholds or a range. 
     
     
         4 . The method of  claim 1 , wherein the one or more predictive models are previously determined from a population of subjects in whom subsequent actual occurrences of the macrophage activation syndrome in a defined time interval were known, and wherein the future time interval is based on the one or more predictive models. 
     
     
         5 . The method of  claim 1  further comprising:
 upon emitting the notification, monitoring at least the one or more first laboratory or physiologic measurements and a diagnosis of the subject; 
 determining, at least partially based on monitoring the one or more first laboratory or physiologic measurements and the diagnosis, a status of the macrophage activation syndrome; 
 generating an updated one or more predictive models based on the status of the macrophage activation syndrome, the one or more first laboratory or physiologic measurements, the diagnosis, and the one or more predictive models; and 
 redetermining the high-frequency components of the power spectrum or the surrogate measure of the high-frequency band of the power spectrum by using the updated one or more predictive models. 
 
     
     
         6 . The method of  claim 1 , further comprising at least one of:
 placing a new orders in an electronic medical record (EMR) of the subject;   modifying a care plan of the subject;   reserving resources, for the subject, in a care facility needed to treat the macrophage activation syndrome;   instituting automatic discharge prevention measures comprising preventing and interrupting discharge orders from being submitted for the subject until such time as the caregiver has reviewed the notification and disabled the automatic discharge prevention measures;   ordering increased monitoring of the subject; and   ordering increased testing of the subject or ordering prescriptions for the subject.   
     
     
         7 . The method of  claim 1 , further comprising:
 monitoring a diagnosis of the subject;   determining, based on the diagnosis, an occurrence or nonoccurrence of the macrophage activation syndrome; and   updating, based on the occurrence of the macrophage activation syndrome, the one or more predictive models.   
     
     
         8 . A system comprising:
 a processor; and   a computer storage medium storing computer-usable instructions that, when used by the processor, cause the processor to perform operations comprising:
 receive one or more measurements of a subject, the one or more measurements comprising:
 one or more first laboratory or physiologic measurements and one or more corresponding date-time coordinates; and 
 one or more second laboratory or physiologic measurements and a respective date-time coordinate, the one or more second laboratory or physiologic measurements being less recent than the one or more date-time coordinates corresponding to the one or more first laboratory or physiologic measurements; 
 
 retrieve a time series including at least the one or more first laboratory or physiologic measurements and the one or more second laboratory or physiologic measurements, the time series based on the one or more corresponding date-time coordinates of the one or more first laboratory or physiologic measurements and the respective date-time coordinate of the one or more second laboratory or physiologic measurements; 
 determine a probability of a macrophage activation syndrome for the subject by determining, for the time series, high-frequency components of a power spectrum or a surrogate measure of a high-frequency band of the power spectrum, the probability determinable by using one or more predictive models; 
 determine that the probability of the macrophage activation syndrome exceeds a threshold associated with a condition of the subject, with the one or more first laboratory or physiologic measurements, with the one or more second laboratory or physiologic measurements, or with other laboratory or physiologic measurements; 
 in response to determining the probability of the macrophage activation syndrome exceeds the threshold, identify a recommendation for a treatment or a therapy for the subject; and 
 transmit a notification to a caregiver indicating that the subject has the probability, corresponding to a future time interval, of the macrophage activation syndrome exceeding the threshold, the notification comprising the recommendation. 
   
     
     
         9 . The system of  claim 8 , wherein the operations further comprise, based on the probability satisfying the threshold, perform an action comprising at least one of:
 placing a new order in an electronic medical record (EMR) of the subject;   altering a care plan of the subject;   reserving a resource, for the subject, in a care facility for treating the macrophage activation syndrome;   modifying a plan of care for the subject that prevents, interrupts, or delays discharge orders from being submitted or carried out;   automatically scheduling increased monitoring of the subject;   ordering increased testing of the subject;   automatically scheduling a consultation with a care provider;   automatically issuing a clinical order for the subject; or   issuing an electronic alert or notification to the care provider or the subject.   
     
     
         10 . The system of  claim 8 , wherein the computer storage medium stores maximum and minimum values defining target ranges for the one or more first laboratory or physiologic measurements. 
     
     
         11 . The system of  claim 8 , wherein the operations further comprise:
 determine a plurality of future values representative of a corresponding plurality of expected levels of the subject, and   cause a display of the plurality of future values in graphical form on a computing device.   
     
     
         12 . The system of  claim 8 , wherein the operations further comprise calculate a Hurst exponent, fractal dimension, or other self-similarity metrics of the at least one time series, and calculate an entropy or chaotic variability of the at least one time series. 
     
     
         13 . The system of  claim 8 , wherein the operations further comprise receive the one or more predictive models that are previously determinable from a population of subjects in whom subsequent actual occurrences of the macrophage activation syndrome in a defined future time interval were known. 
     
     
         14 . The system of  claim 8 , wherein the one or more first laboratory or physiologic measurements comprise at least one of levels of: ferritin, triglycerides, fibrinogen, erythrocytes, leukocytes, platelets, albumin, lactate dehydrogenase (LDH), creatinine, resting respiratory rate, resting heart rate, systolic blood pressure, or body temperature. 
     
     
         15 . The system of  claim 14 , wherein the at least one of the levels of the LDH are from a sample of the subject, and wherein a serial serum level of the LDH is 250 U/L or greater. 
     
     
         16 . The system of  claim 14 , wherein a serial serum level of the ferritin in a sample of the subject is 150 ng/mL or greater. 
     
     
         17 . The system of  claim 14 , wherein a serial serum level of the fibrinogen is 200 mg/dL or less, and wherein a level of the triglycerides is 180 mg/dL or greater. 
     
     
         18 . The system of  claim 14 , wherein:
 RMSSD3 (LDH) is greater than 12 U/L/day, at least one LDH level included in a calculation is greater than 250 U/L;   RMSSD3 (ferritin) is greater than 15 ng/mL/day, and at least one Ferritin level included in the calculation is greater than 150 ng/mL;   RMSSD3 (fibrinogen) is greater than 10 mg/dL/day, and at least one Fibrinogen level included in the calculation is less than 200 mg/dL;   RMSSD3 (triglycerides) is greater than 18 mg/dL/day, and at least one Triglycerides level included in the calculation is greater than 180 mg/dL; and   wherein specimens utilized for determination of the levels and RMSSD3 values are separated by not less than 24 hours and not more than 30 days.   
     
     
         19 . A non-transitory computer-readable medium comprising instructions executable by a processor that, when executed, cause the processor to perform operations comprising:
 receiving one or more measurements of a subject, the one or more measurements comprising:
 one or more first laboratory or physiologic measurements and one or more corresponding date-time coordinates; and 
 one or more second laboratory or physiologic measurements and a respective date-time coordinate, the one or more second laboratory or physiologic measurements being less recent than the one or more date-time coordinates corresponding to the one or more first laboratory or physiologic measurements; 
   retrieving a time series including at least the one or more first laboratory or physiologic measurements and the one or more second laboratory or physiologic measurements, the time series based on the one or more corresponding date-time coordinates of the one or more first laboratory or physiologic measurements and the respective date-time coordinate of the one or more second laboratory or physiologic measurements;   determining a probability of a macrophage activation syndrome for the subject by determining, for the time series, high-frequency components of a power spectrum or a surrogate measure of a high-frequency band of the power spectrum, the probability determinable by using one or more predictive models;   determining that the probability of the macrophage activation syndrome exceeds a threshold associated with a condition of the subject, with the one or more first laboratory or physiologic measurements, with the one or more second laboratory or physiologic measurements, or with other laboratory or physiologic measurements;   in response to determining the probability of the macrophage activation syndrome exceeds the threshold, identifying a recommendation for a treatment or a therapy for the subject; and   transmitting a notification to a caregiver indicating that the subject has the probability, corresponding to a future time interval, of the macrophage activation syndrome exceeding the threshold, the notification comprising the recommendation.   
     
     
         20 . The non-transitory computer-readable medium of  claim 19 , wherein the probability satisfies the one or more thresholds when the probability equals or exceeds the one or more thresholds.