Data analysis mechanism for generating statistics, reports and measurements for healthcare decisions

Abstract

A system obtains patient data, de-identifies each patient with a unique patient key (UPK) and joins the patient data without linking to personal identities of the patients. The system further receives a search query from a user device, identifies a group of UPKs associated with patient data that satisfies criteria of the search query based on values for associated healthcare attributes, and clusters the group of UPKs with respect to healthcare events. The system then ranks the healthcare events of the clusters and sends a query result including healthcare data that is generated based on the ranking, where the query result conceals any personal identity of any of the patients.

What is claimed is: 
     
         1 . At least one computer-readable medium, excluding transitory signals and carrying instructions, which, when executed by at least one data processor of a system, cause the system to:
 obtain patient data of multiple patients,
 wherein the patient data includes values for multiple healthcare attributes and descriptions of multiple healthcare events, 
   de-identify each of the multiple patients of the patient data with a unique patient key (UPK);   join the patient data without linking to a personal identity of any of the multiple patients;   receive, over a computer network, a search query from a user device,
 wherein the search query includes criteria for the multiple healthcare attributes; 
   identify a group of UPKs associated with patient data that satisfies the criteria based on the values for the multiple healthcare attributes;   cluster the group of UPKs into multiple clusters with respect to the multiple healthcare events;   rank the multiple healthcare events of the multiple clusters based on an occurrence of each healthcare event;   send, over the computer network to the user device, a query result including healthcare data that is generated based on the ranking of the multiple healthcare events; and   cause display, on the user device, of the query result while concealing any personal identity of any of the multiple patients.   
     
     
         2 . The computer-readable medium of  claim 1 , wherein the multiple healthcare events are ranked based on a frequency of the occurrence of each healthcare event. 
     
     
         3 . The computer-readable medium of  claim 1 , wherein the ranking of the multiple healthcare events is based on occurrences of individual healthcare events and co-occurrences of two or more healthcare events. 
     
     
         4 . The computer-readable medium of  claim 1 , wherein each description of a healthcare event includes a weight associated with a respective UPK, and wherein the system is further caused to:
 determine a subset of the multiple healthcare events for the group of UPKs based on associated weights,   wherein the clustering of the group of UPKs is performed with respect to the subset of healthcare events.   
     
     
         5 . The computer-readable medium of  claim 4 , wherein the weight is related to a significance of a respective healthcare event to a patient associated with the respective UPK. 
     
     
         6 . The computer-readable medium of  claim 4 , wherein the weight is related to a quantity of times that a respective healthcare event occurred to a patient associated with the respective UPK. 
     
     
         7 . The computer-readable medium of  claim 4 , wherein to determine the subset of the multiple healthcare events comprises causing the system to:
 select a specific number of healthcare events with largest sums of weights over the group of patients.   
     
     
         8 . The computer-readable medium of  claim 4 , wherein the subset of the multiple healthcare events is determined based on a principal component analysis. 
     
     
         9 . The computer-readable medium of  claim 1 , wherein the system is further caused to:
 send, over the computer network to the user device, an indication of availability of alternative treatment plans, which are related to the clusters and corresponding rankings of the multiple healthcare events.   
     
     
         10 . The computer-readable medium of  claim 1 , wherein the multiple healthcare attributes include demographic information, location information, or medical information of the multiple patients. 
     
     
         11 . The computer-readable medium of  claim 10 , wherein the medical information includes comorbidities of the multiple patients. 
     
     
         12 . The computer-readable medium of  claim 1 , wherein the multiple healthcare events are associated with healthcare providers, healthcare facilities, or healthcare procedures. 
     
     
         13 . The computer-readable medium of  claim 1 , wherein the multiple healthcare events are associated with cost, and wherein the system is further caused to:
 compute a sum of costs associated with the multiple healthcare events for each UPK in a highest-ranking cluster of UPKs,   wherein the query result includes an aggregate or a distribution of the sums.   
     
     
         14 . The computer-readable medium of  claim 1 , wherein the multiple healthcare attributes include a healthcare provider, and wherein the system is further caused to:
 for a highest-ranking cluster of UPKs, identify a group of healthcare providers associated with the multiple healthcare events,   wherein the query result includes at least part of the group of healthcare providers.   
     
     
         15 . The computer-readable medium of  claim 14 , wherein the system is further caused to:
 for the highest-ranking cluster of UPKs, extend the group of healthcare providers with additional healthcare providers affiliated with the group of healthcare providers,   wherein an affiliation is based on patient referral, patient sharing, or facility sharing, and   wherein the query result includes the extended group of healthcare providers.   
     
     
         16 . The computer-readable medium of  claim 1 , wherein the multiple healthcare attributes include a facility type, and wherein the system is further caused to:
 for a highest-ranking cluster of UPKs, identify a group of facilities associated with the multiple healthcare events,   wherein the query result includes at least part of the group of facilities.   
     
     
         17 . A system comprising:
 a processor; and   a memory storing instructions which, when executed by the processor, cause the system to:
 de-identify any patient identified in patient data with a unique patient key (UPK),
 wherein the patient data includes values for multiple healthcare attributes and descriptions of multiple healthcare events; 
 
 receive, over a computer network, a search query from a user device,
 wherein the search query includes criteria for the multiple healthcare attributes; 
 
 identify a group of UPKs associated with patient data that satisfies the criteria based on the values for the multiple healthcare attributes; 
 cluster the group of UPKs into multiple clusters with respect to the multiple healthcare events; 
 rank the multiple healthcare events of the multiple clusters based on an occurrence of each healthcare event; 
 send, over the computer network to the user device, a query result including healthcare data that is generated based on the ranking of the multiple healthcare events; and 
 cause display, on the user device, of the query result while concealing any personal identity of any of the multiple patients. 
   
     
     
         18 . The system of  claim 17 , wherein each description includes a weight associated with a respective UPK, and wherein the system is further caused to:
 determine a subset of the multiple healthcare events for the group of UPKs based on associated weights,   wherein the clustering of the group of UPKs is performed with respect to the subset of healthcare events.   
     
     
         19 . A method comprising:
 receiving, over a computer network, a search query from a user device,
 wherein the search query includes criteria for multiple healthcare attributes included in patient data, 
 wherein the patient data additionally includes descriptions of multiple healthcare events, and 
 wherein the patient data de-identifies any patient with a unique patient key (UPK); 
   identifying a group of UPKs associated with patient data that satisfies the criteria based on values for the multiple healthcare attributes;   clustering the group of UPKs into multiple clusters with respect to the multiple healthcare events;   ranking the multiple healthcare events of the multiple clusters based on an occurrence of each healthcare event;   sending, over the computer network to the user device, a query result including healthcare data that is generated based on the ranking of the multiple healthcare events; and   causing display, on the user device, of the query result while concealing any personal identity of any patient of the patient data.   
     
     
         20 . The method of  claim 19  further comprising, prior to receiving the search query:
 obtaining the patient data of multiple patients; 
 de-identifying each of the multiple patients of the patient data with a respective UPK; and 
 joining the patient data without linking to a personal identity of any of the multiple patients.