Correlation of clinical events

FIELD: medicine.

SUBSTANCE: invention refers to event correlation systems. A health care facility network contains a number of electronically stored patient records comprising electronically searchable data, a base event list consisting of a number of base event definitions, a related event list consisting of a number of related event definitions that can be correlated with the base events, and a correlation processor that uses the base event definition and at least one related event definition to relate the events and scans the patient records to search a pre-assigned correlation.

EFFECT: enhancing the collection and correlation of the events providing better follow-up and clinical results in a patient managed by a doctor.

21 cl, 5 dwg

The present invention relates to patient care and treatment in a clinical setting. In particular, the present invention finds use in determining the correlation of events that occur during patient care, and will be described with particular reference to the above-mentioned care. It should be understood that the present invention can be applied in any situation in which events are logged, and not necessarily only in terms of clinical care.

Data written to the card of the patient, can be viewed as a sequence of events in time. These events are often associated with other events, which are also recorded in the card. Retrospective analysis of information presents a number of problems. First, the data change over time. As patient care in the patient continuously add new measurements, observations, summaries, diagnoses, instructions, etc. of the Relationships between data elements are also changed, for example, on the patient condition as appropriate therapeutic measures or Protocol, executable in a clinical setting, and other Events that may be irrelevant in the care of one patient, may be relevant to the diagnosis of the patient or care depending on the individual circumstances of the patient.

Previous tools for data analysis was focused on financial and operational segments of commercial enterprises. The mentioned tools are typically focuses mainly on data that can easily be reduced to summarize, to sum or count. In contrast to the analysis of financial data analysis of clinical data usually focuses more on identifying the set of physiological conditions and the determination of the existence and influence of concomitant medical interventions and care. The mentioned type of data changes over time and the relationship of one event to another can vary depending on the patient's condition and perform therapeutic measures.

Commercial tools usually are not able to work in such analysis, which focuses directly on the main task of the doctor, and the time to improve care and clinical outcomes for patients of the doctor. For example, it is assumed that patients, mean arterial pressure which falls below 65, should be treated vasoconstrictor drugs bolus dose of fluid or other common method of treatment. It is also expected that the treatment should be carried out within a certain time interval of observation events low blood pressure. Currently there is no reliable way, which allows physicians to evaluate whether the treatment of the aforementioned type consistent for all patients and with equal attention during all the day.

The present application provides new and improved method and apparatus for accumulation and correlation of important events in the course of patient care, which overcome the above and other problems.

In accordance with one aspect proposes a method of establishing a correlation of the observed events in the patient records. Get the definition of basic events, which reflects the observation of events during the patient's stay in a medical facility, which is recorded in the card patients. Receiving a definition of at least one associated event, which is observed in the cards containing the basic event. Get the definition of at least one relationship, at least, one of the CSOs associated event basic event. Card patients are looking to search for cards containing the basic event and the associated event is related by definition. Generate a report that explains the observation of basic events with at least one associated event, as defined, of at least one relationship.

In accordance with another aspect of the proposed network of medical institutions. The network includes many patient records stored in electronic memory, this card contains data that are available for electronic search. The list of basic events contains many definitions of basic events. A list of related events contains many definitions related events, for which it is possible to establish a correlation with the underlying event. The correlation engine uses the definitions of the basic events and at least one associated event and given the definition of the relationship between events and looking at the card patients to search for correlation according to the task.

In accordance with another aspect features a method for detecting correlations of events. Choose at least one definition of the basic events. Choose at least one definition of the relationship with the underlying event. Looking at the card patients to search for related events that satisfy usamos the connection with each set defining the underlying event. Produces a report that represents the detected events to the user.

One advantage is to establish the correlation of different data blocks in the patient card.

Another advantage is the control for the wrong patient care.

Another advantage is to establish the correlation data in the patient with the patient's response to therapeutic measures.

Another advantage is the possibility of joint use of correlations.

Another benefit is in helping to identify and eliminate the unfair practice of medicine.

Additional advantages of the present invention will be obvious to specialists with an average level of competence in the art after reading and studying the following detailed description.

The invention can be performed in the form of the various components and combinations of components, and in various steps and combinations of steps. The drawings are intended only to illustrate preferred embodiments and do not allow their interpretation in the sense of limiting the invention.

Figure 1 - schematic of the system to establish the correlation of events in accordance with the present application;

figure 2 - block diagram of a sequence of exemplary steps performed in the process set the effect to correlate events;

figure 3 - example of a summary report for presentation to the user;

figure 4 - example of a detailed report for presentation to the user;

figure 5 - example of a summary schedule for presentation to the user.

As shown in figure 1, the patient is10is clinical conditions and receives long-term care. During the course of patient care performed various measurements relating to the health of the patient. The above measurements can be regular, for example, measurements of blood pressure, pulse rate, body temperature, blood sugar levels, etc. or the measurement may be less regular, for example ECG, electrocardiogram taken during one of the exercise load, etc. Measurements can be performed automatically by sensors12located on the patient, and recorded by the monitor14the patient, or the measurements can be performed manually by a medical specialist such as a nurse, doctor, hospital attendant or technician. Measurements can also be performed in the laboratory tests.

Routinely mentioned measurements are supplied with a time stamp and stored in the card16patient. In a preferred embodiment, the card is electronic and available to medical professionals with the appropriate category of admission is for the network 18medical institutions. In a medical facility can continue to use paper cards and record measurements by hand. In this situation, the medical specialist must then manually enter the information in the electronic map16a patient with a computer20connected to the network (for example, post nurses), or any other wireless portable devices22connected to the network18institutions, for example, a tablet personal computer, portable computer, handheld computer, Imagination, Blackberry, cell phone, etc. in Addition, the network18does not have to be limited to one medical institution; the network may contain multiple institutions or even the database for General use (without identifiers of patients to privacy policy).

Medical assistants will be planned in order to enter other information in the patient. Referred to input, not inherently limited in scope and may include comments received during the collection of patient history, mental status, pallor, estimated diagnoses, care instructions, taken drugs or therapeutic measures, test results, observations, events, clinical consultations, and many other features that may have an IU is in a period of stay of a patient in a medical facility. In a relatively short time card16the patient may become very lengthy.

Online18institutions can also be stored card other patients (16a,16b,...,16n). The elements contained in the patient records, can often be of interest to a medical professional, such as shift supervisor, who is interested in improving the efficiency of its staff. Perhaps the doctor reads an article in a medical journal and intends to establish, apply whether his institution methods proposed in the paper. Perhaps in-house lawyer learned about any other institution conflict with the law because of some methods of medical practice. Then the lawyer can verify that similar methods of medical practices are not applied in the institution, which is referred to the lawyer. In this case, it is useful to define the correlation of key events, which are included in the card patients.

When the content is too much data in the patient records manual viewing information can be impractical and prone to subjective errors. In addition, the observation of a single event is often not of interest. Several related events are often more informed about how well the institution. For example, if mean arterial pressure (MAP) of the patient drops below 65, is in accordance with common practice to deliver a bolus dose of intravenous fluids and/or treatment vasoconstrictor drugs. If the medical specialist can install a correlation of one event with another, namely how often have low blood pressure can be cured within a reasonable time, then the above-mentioned correlation has a high value for the investigator. Network18medical institutions contains the correlation processor24that allows medical specialist on the user interface to Express detection criteria and the relationship between one event and another event and identify the event in the time series data. Medical professionals can retrospectively to view a group of patients and to determine when or how often have at least two clinical events and the relationship between them in time.

The figure 2 shows, with reference to figure 1 block diagram of the operational sequence of the method in accordance with an exemplary embodiment. As mentioned above, card patients filled the stage30measurements, annotations, instructions, diagnosis, comments, etc. or automatically measuring devices12or from the hands of medical professionals. When the medical specialist is ready to establish the correlation of events, i.e. has the idea for the study, the user accesses the correlation p is accessory on stage 32. The above step can be performed using a graphical user interface contained in the composition of the network computer20or any other portable device22. Depending on the correlation, which intends to establish a medical specialist, he may wish to set boundaries volume correlation. For example, the specialist may wish to explore the viability of some treatment of the disease, which means search all available cards. In an alternative embodiment, the specialist may wish to obtain information relating to the care provided to specific therapeutic unit that, from the outset, limits the information required by the much smaller subset of the entire group of patients. In this regard, the user may optionally first be defined subset of patients at the stage of34although the user can view all the cards in the database according to his desire. To ensure the choice of some parameters of the patient based on the location of the patient, date of implementation of the care, the demographic parameters of the patient, type of care/hospitalization, result, etc., apply a set of filters that36. You can use any combination of the aforementioned or other filters, and the choices of filters can be combined using l the logical operators (for example, And, OR) to form intersections between filters. Exemplary non-exhaustive list of possible filters contains clinical unit, Department, type of admission, date of care, mortality, place of discharge, hospital care, the root cause of hospitalization, the patient's age, date of birth, ethnic group, nationality, type of patient and the race. After the desired(s) filter(s)36selected filter36acts on stage38to exclude card patients, not included in the user request. In one example, the user may wish to filter all patients, except those who are male, hospitalized in the intensive care unit and was hospitalized for a user-specified two-month period.

After the search filters are applied, the user specifies the definition of the base events on stage40. The underlying event is a primary event with which you want to find the correlation of other events. The user selects the underlying event at the stage of40from the list that is automatically generated. In an alternative embodiment, the user can specifically set the basic event. The list contains all the data recorded in the card in the medical information system, and is composed of all data elements and their attributes. In the can, the variability and diversity of language may delay the process at this stage. For example, if the user chose "heart attack" as a basic event, but many other medical experts called this event "myocardial infarction" or the abbreviation "MI" when recording events in the map, the user may indiscretion to lose valuable data. In such a situation, useful Systematized nomenclature of medical terms or language system "SNOMED", as it standardizes medical professional language. The system uses SNOMED common identifiers to reduce the likelihood that the corresponding data will be lost due to different language variants. Another possible applicable system is a system ICD9 that standardizes codes for billing. If the underlying event involves some sort of account statement, the system ICD9 can support the use of professional language, billing as well as system SNOMED can support the use of professional medical language.

With regard to the General patient selection, the user may wish to classify the underlying event is further through the use of filters on stage44. You can use any combination of the aforementioned filters and similarly shown that the am band filters basic events can be combined using logical operators. Selectable filters can be pre-defined or based on the properties of the selected data. Exemplary non-exhaustive list of properties includes a numeric, string and date value, unit of measure, the relevant material in the appropriate location, current location, time, memory, and time, recorded in the card. Exemplary non-exhaustive list of filters includes statements such as there is, equal to, less than, less than, or equal to, greater than, greater than or equal increases for at least "x" period of time "y"is reduced by at least "x" period of time "y", similarly, is the minimum value, the maximum value is the first event recorded in the card, and is the last event recorded in the card. When you call for example blood pressure selected basic event may contain MAP (mean arterial pressure) below 65, which declined during the two-hour period on at least 5 mm RT. Art. in Addition, the user can select specific values for the properties, get back with the data.

Then, the user specifies at least one related event for which you want to establish a correlation with the underlying event, on stage46 . Similarly, the basic events related events can be stored in the database48related events in the form of descriptions in standardized terminology, SNOMED. Also similar to the basic events, the user can classify related events by applying filters on stage50to further restrict the data that will come with the answer. Then, the user can also specify the relationship that exists between each related event and the selected basic event, on stage52. The user can set the above relationship as a function of time (for example, within "x" minutes for the underlying event) or through at least one relationship between the properties of the underlying events and properties associated event.

Upon completion of the drafting of the definition of correlation, the user can save the correlation at the stage of54in the correlation memory55. It is assumed that the user wishes to conduct a correlation immediately, but this is optional. In addition, the correlation can how to spend and save in memory with the intention to carry out the correlation again later or periodically. For example, if shift correlates immediately and detects a fault, the shift supervisor may take action to eliminate the aforementioned drawback. Several weeks the representatives later shift may perform the correlation again to assess, did measure the desired effect on events of interest.

When the user carries out the correlation, a report is generated at the stage of56. The user can view the results on the screen, print the results, set the correlation to write its results to the database of medical information systems, schedule continuous analysis of patient records using the correlation or publish correlation, so that other users can use it. The figure 3 shows a sample summary report60that may be a correlation processor. In the summary report presents summary data correlation. Optionally, as shown in figure 4, the user can generate a more detailed summary62that can show individual results, which are compiled in a summary report60. In addition, the user may select a graphical representation of their data. As shown in figure 5, the user generates a graph64that compares attention to the time of day. On the x-axis presents the time of day and the y-axis reflects the percentage of events that require the attention of a medical professional and provided assistance in the prescribed time.

It should be understood that the user can create is your own basic events or related events and is not limited by language or SNOMED events contained in the list42basic events or the list of48related events. Medicine moves forward, and as new diagnoses and new treatments, users will be freed from the restrictions of old definitions or old treatment measures or not will have to wait for the software update, which includes new data. If the description does not yet exist, you can create a description that satisfies its requirements in this situation.

In addition, significant correlation can be held in memory55and register it for use without the need to create. Data is tested and known to give satisfactory results. In the above-mentioned correlations to the user, at least, will not need to worry about whether adequate he described the correlation (e.g., filters out whether the request too many cases was not whether the request is overly broad, and the like). Some rough correlation is described below:

Correlation to drug treatment

• The appointment of vasoconstrictor drugs, depending on the MAP (mean arterial pressure) of the patient

• Insulin depending on the level of glucose in the blood serum of the patient

• Purpose of propofol depending on the depth of the Glasgow coma

• The appointment of morphine or fentanyl, depending on the indicator of pain the patient

• Administration of diuretics, depending on the pressure in the pulmonary artery (PAP) patient data (SpO₂) pulse oximetry and diuresis

• Purpose of nitroprusside depending on MAP patient and intracranial pressure (ICP)

Correlations for liquids

• Intravenous bolus dose of fluid depending on Central venous pressure (CVP) of the patient, PAP and MAP

• Delivery of erythrocyte mass, depending on the hematocrit (HCT) of the patient, oxygen (O₂Sat) and partial pressure of arterial oxygen (PAO₂)

• Delivery of platelets depending on the number of platelets of the patient

• Delivery of total parenteral nutrition (TPN), depending on the patient's blood glucose

• Shipping colloids depending on serum albumin patient

Diagnostic correlation

• Diagnosis of severe sepsis, depending on the previous number of leukocytes (WBC), temperature and blood pressure of the patient, and the duration of the patient's stay and mortality

• The diagnosis of acute respiratory distress syndrome depending on the previous total volume of the patient, mode of ventilation with continuous positive pressure at the end of exhalation and the WBC/p>

• The diagnosis of hypovolemia depending on the total volume of fluid in the patient

• Diagnosis of renal failure depending on previous levels of creatinine and nitrogen in urea patient's blood

In an alternative embodiment, the user must select the associated event with which you want to define the correlation of the underlying events. This alternative implementation is useful from a research point of view, and the correlation processor 24 is used to retrieve information from a data set correlations instead of searching the user-selected correlations. A medical expert will use this alternative implementation, when there is a base event about which medical specialist wants to find out more. In the illustrative example, the medical officer notices an abnormally high frequency of postoperative infection. When trying to determine the cause of the infection specialist performs a search for any event that occurred a day before the onset of infection in at least 90% of patients with emerging infection. Many of issued correlations may be dismissed as random, but the specialist may encounter on the overall event, which will explain observations of events infections.

The invention is described with reference to preferred options for implementation. Specially the East will be able to create modifications and changes after reading and understanding the foregoing detailed description. Interpretation invention provides coverage of all of the above modifications and changes to such an extent that they are within the scope of the claims appended claims or their equivalents.

1. The network of medical institutions, containing:
many cards (16) patients, stored in the electronic memory, this card contains data that are available for electronic retrieval system;
the list (42) of the underlying event, which contains many definitions of basic events;
the list (48) related events, which contains a number of definitions related events, for which it is possible to establish a correlation with the underlying event;
the correlation processor (24), which uses the definition of the underlying events and at least one associated event and given the definition of the relationship between events and looking at the card patients to search for correlation according to the task.

2. The network of medical institutions according to claim 1, additionally containing
filters (36), which act as a preliminary filter out unwanted results, as specified by the user.

3. The network of medical institutions according to claim 1, additionally containing a correlation memory (55), which stores previously specified correlation for subsequent use or distribution.

4. The network is MEDICINSKOGO institution according to claim 1, in which the list (42, 48) of events normalized descriptions system SNOMED medical language.

5. The network of medical institutions according to claim 1, additionally containing a user interface (20, 22), which introduce the relationship between events using at least one Boolean operator.

6. The network of medical institutions according to claim 1, in which the correlation processor (24) generates a report, which is correlated to the user.

7. The way to establish a correlation of the observed events in the patient records, the method includes the steps are:
get the definition of basic events, which reflects the observation of events during the patient's stay in a medical facility, which is recorded in cards of patients;
receiving a definition of at least one associated event, which is observed in the cards containing the basic event;
get the definition of at least one relationship, at least one associated event basic event;
looking at the card patients to search for cards containing the basic event and the associated event is related by definition; and
generate a report that explains the observation of basic events with at least one associated event, as defined, of at least one relationship.

8. The way p is 7, optionally containing phases in which:
get the filter definition;
filtered subset of patient records that should not be performed, as described in the filter definition.

9. The method according to claim 7, additionally containing a stage, on which:
retain the definition of events and the relationship to subsequent correlation.

10. The method according to claim 9, further containing a stage, on which:
distribute definitions are saved and the relationship among medical professionals to use them.

11. The method according to claim 7, in which the step of obtaining basic event contains the stage at which:
the user selects the basic event basic event.

12. The method according to claim 11, in which the list of basic events normalized descriptions system SNOMED medical language.

13. The method according to claim 7, in which the received definition systematically generated electronically for finding relationships.

14. The method according to claim 7, in which at least one relationship described at least one Boolean operator.

15. The method according to claim 7, in which the step of generating the report includes a step of forming a report that gives a summary of the correlation data.

16. The method according to claim 7, in which the step of generating the report includes a step of forming a detailed list of data correlation.

17. The method according to claim 7, in which et is p report includes a step of forming a graphical representation of the correlation data.

18. The method according to claim 7, in which at the stage of report generation is formed, at least one of:
summary report;
detailed report;
graphical report;
display report;
a printed report;
report stored in the database system;
publishing correlation; and,
the number of repeated cases of correlation.

19. A computer storage medium, which performs the method according to claim 7.

20. The way to detect correlations of events, the method includes the steps are:
choose at least one definition of the underlying event;
choose at least one definition of the relationship with the underlying event;
looking at the card (16) patients to search for related events that satisfy relationships with each set defining the underlying event;
produces a report that represents the detected events to the user.

21. The method according to claim 20, in which the generated report contains the percentage of observations of events in the cards (16), containing the basic event and the selected relationship.