System for determining a patient's risk of falling

Abstract

Title of the invention System for determining a risk of falling in a patient The invention relates to a system (1) for determining a risk of falling in a patient (P) comprising means for measuring the weight (2) of the patient; means for collecting and remotely transmitting patient data (3) for transmitting a patient data set (Edp) comprising the measured weight (Po), a patient identification (IDp), a time identifier (IDt) and at least one or more events (Ev) linked to the patient, and at least one remote server (4) comprising reception means configured to receive the patient data sets (Edp). The server (4) makes it possible to store (4.1) and process (4.2) the patient data sets (Edp) to determine a fall risk score from the calculated indicators, and display means (5) for displaying fall risk information relating to the fall risk score determined for the patient in question.Abstract Figure: Figure 1.

Description

Title of the invention: System for determining a risk of falling in a patient

[0001] The invention relates to a system for determining a risk of falling in a patient, in particular a geriatric patient at high risk of falling, who is "hospitalized" at home.

[0002] Falls in geriatric patients, i.e. elderly people, are known to have a negative impact on their quality of life due to the risk of injury, in particular a risk of losing their ability to walk if, after their fall, these patients remain lying down for too long without assistance.

[0003] In an attempt to remedy this, alert devices have been proposed that patients activate in the event of a fall. However, in practice, it is observed that some patients often forget to bring them with them, and therefore cannot trigger an alert in the event of a fall. Furthermore, although they can provide an alert in the event of a fall, they do not prevent or predict a fall, and therefore do not prevent injuries caused by falls and the associated consequences.

[0004] It would be significantly more beneficial for patients to be able to prevent their falls so that they can be avoided or reduced in frequency.

[0005] The document Mishra, AK, et al.; Explainable Fail Risk Prediction in Older Adults Using Gait and Geriatric Assessments; In Frontiers in Digital Health (Vol. 4); 2022; Frontiers Media SA, proposes to prevent the risk of patients falling by algorithmically processing different patient parameters, such as FAP (functional ambulation score), ambulation speed, GDS score (geriatric depression score), MMSE score (cognitive function assessment test), IADL (score related to the need for assistance in daily activities), etc.

[0006] However, the proposed solution is not ideal because the described approach is limited by the need to measure all of these descriptors in their entirety.

[0007] and requires for some specific equipment which is not necessarily available, in particular at the patient's home, such as the treadmill system required to measure walking speed.

[0008] A problem is therefore to be able to improve the prediction of the risk of falling in a patient (or patients) at high risk of falling, in particular a geriatric patient, who is “hospitalized” at home.

[0009] A solution according to the invention relates to a system for determining a risk of falling in a patient comprising: - means of measuring the patient's weight to determine the patient's weight during visits by a healthcare professional taking place on several consecutive or non-consecutive days, - means for collecting and transmitting patient data to transmit, during each visit, at least one set of data including the measured weight (po), a patient identification (IDp), a time identifier (IDt) and one or more events (Ev) linked to the patient, and - at least one remote server with receiving means configured to receive the patient data sets, wherein said at least one server is configured to: i. store the received patient data sets and ii. process patient data sets by: a. a calculation of several predefined indicators from the stored data sets, each indicator reflecting a history of one of the patient-related events occurring during a fixed elapsed period of time, and b. a determination of a fall risk score from the calculated indicators, - display means cooperating with said at least one remote server to display fall risk information relating to the determined fall risk score.

[0010] Depending on the embodiment considered, the system of the invention may comprise one or more of the following characteristics: - said at least one remote server (i.e. computer server(s)) comprises a data storage server and a calculation server. - said at least one remote server comprises one or more Cloud or Edge type servers installed near the data sources. - the means of collecting and transmitting patient data include a portable device such as a laptop, digital tablet or multifunction telephone (i.e. smartphone), preferably a digital tablet. - the means for collecting and transmitting patient data are configured to transmit said at least one set of patient data via a low-energy GSM (4G / 5G) or Sigfox, Lora type telecommunications protocol, etc. the means of collecting and transmitting patient data are located at the patient's home. patient-related events (Ev) include data relating to: • one or more medical scores, including the Conley score, which serves as a predictor of the risk of falling; the Wong-Baker Faces Scale score, which standardizes the measurement of the patient's pain level; the IADL score, which measures a patient's level of independence in carrying out daily living tasks; the Braden score, which measures the risk of developing pressure sores during hospital care, etc., • the existence of significant pathologies or treatments, in particular antibiotic therapy, gastric tube, urinary tube, respirator, etc. • the presence of particular risk factors, including unsuitable housing, difficulty getting around, advanced age, absence of a family carer, etc., • the presence of paramedical treatments, in particular physiotherapy, speech therapy, etc., • patient physiological data, including weight / BMI, vital signs, etc., and / or • sociodemographic data, including place of residence, age, sex, presence of a caregiver, etc. at least a portion of the events (Ev) related to the patient are entered into the portable device, preferably a digital tablet, by a healthcare worker visiting the patient at home, such as a nurse or the like. patient-related events (Ev) are collected over a fixed elapsed time period of several months, typically 2 to 12 months, for example of the order of 6 months. the system is configured to determine the fall risk score(s) of the patient under consideration over a period of several weeks or several months, preferably within the next 3 months. the system is configured to store the fall risk score(s) of a given patient, i.e. a history of the fall risk scores of the patient in question, or of several patients. the system is configured to display, for a given patient, the successive fall risk scores having been determined over a given period of time, for example several months, for the patient in question, i.e. to display an assessment over time of the patient's fall risk. said at least one server, in particular a calculation server, is configured to process the patient data sets (Edp) by performing a calculation of several predefined indicators comprising: • for Boolean values: the existence of at least one event of each type; a number of events of each type; and / or a frequency of events of each type, for example any act performed on the patient during visits, i.e. at each visit or not, • for numerical values: a calculation of the average value of measurements taken, for example weight measurements; a calculation of the last known value of the measurement(s) taken; a calculation of the trend (i.e. increase, decrease or stability) over a given period of time, typically over several months, for example over 6 months; and / or an imputation of any missing value when the value has not been measured, using the value that would be measured in a healthy patient (for example: 22 for the body mass index, 0 for the Conley score, etc.). the display means comprise a display screen, for example the screen of a desktop or laptop computer. the display means, for example the desktop or laptop computer, are located in a treatment planning center. The system is configured to simultaneously determine the risk of falls in several patients, i.e. it allows monitoring and follow-up of several patients, each located at home. the system is configured to propose and display the means of displaying, in the event of determining a high risk of falling in a given patient, a corrective action to be taken, such as a visit from a nursing staff member, a technician intervention or even a medical consultation. the display means are configured to simultaneously display a graphical representation of the risk of falling for several patients, that is to say for the different patients monitored. The graphical representation includes a bar chart, with each bar in the chart corresponding to each patient's fall risk. The graphical representation includes different colors depending on the determined fall risk score (i.e. depending on the severity), for example a green color for a low risk score, a red color for a high risk score and an orange color for an intermediate risk score. - the graphical representation includes patient identifications (e.g. their name or IDp) displayed next to the graph bars. - the graphical representation includes a ranking of patients according to their fall risk score, for example from the patient with the highest risk to the patient with the lowest risk. - the display means are further configured to display, for a given patient P, a graphical representation of the evolution of the fall risk score of this patient over a given period of time, for example several weeks or several months. - the display means are configured to display fall risk information including a numerical fall risk value, such as a percentage (%) and possibly one or more associated causes.

[0011] The invention will now be better understood thanks to the following detailed description, given for illustrative but non-limiting purposes, with reference to the appended figures among which:

[0012] [Fig.l] schematizes an embodiment of a system for determining a risk of falling in a patient according to the invention.

[0013] [Fig.2] schematizes a graphical representation of the fall risk scores determined for several patients by means of a system according to the invention, such as the system of [Fig.l].

[0014] [Fig.3] schematizes a graphic representation of the evolution of the risk of falling determined for a given patient over time, via a system according to the invention, such as the system of [Fig.l].

[0015] [Fig.4] schematizes a graphic representation of the factors which have led to an increase or decrease in the risk of falling for a particular patient, via a system according to the invention, such as the system of [Fig.l].

[0016] [Fig.l] schematizes an embodiment of a system 1 for determining a risk of falling in a patient P according to the invention.

[0017] It comprises means for measuring the weight 2 of the patient to determine the weight Po of the patient during visits by a healthcare professional PS, such as a nurse or the like, taking place on several consecutive or non-consecutive days, i.e. during visits to the home of the patient P.

[0018] The weight measuring means 2 comprise a scale or the like, preferably connected and configured to store and / or transmit the stored weights to patient data collection and remote transmission means 3, such as a digital tablet, a laptop, a multifunction telephone (i.e. a smartphone) or the like.

[0019] The means for collecting and remotely transmitting patient data 3 are configured to transmit, i.e., to remotely transmit, during each visit of the healthcare personnel PS, at least one set of patient data Edp to at least one remote server 4, typically a cloud storage server 4.1.

[0020] The remote transmission 6 of the patient data 3 is done via a GSM protocol (i.e. in 4G or 5G) or similar.

[0021] The patient data set Edp comprises in particular the measured weight Po, a patient identification IDp, a time identifier IDt and one or more events Ev linked to the patient.

[0022] The Ev events include data relating to one or more medical scores, in particular the Conley score which is a predictor of the risk of falling; the Wong-Baker Faces Scale score which makes it possible to standardize the measurement of the patient's pain level; the IADL score which makes it possible to measure the level of independence of a patient in carrying out daily life tasks; the Braden score which makes it possible to measure the risk of developing pressure sores during hospital care, etc.); the existence of notable pathologies or treatments (e.g. antibiotic therapy, gastric tube, urinary catheter, respirator, etc.); the presence of particular risk factors (e.g. unsuitable housing, difficulty in moving around, advanced age, absence of a family caregiver, etc.); and / or the presence of paramedical treatments (e.g. physiotherapy, speech therapy, etc.)...

[0023] The remote server(s) 4 comprises, for its part, reception means configured to receive the Edp patient data sets. The remote server 4 is typically a so-called “cloud” server, i.e. a virtual or physical server hosted remotely, deployed on a network and which clients access via a connection (typically the internet).

[0024] Preferably, the server 4 comprises a data storage (sub-)server 4.1 for receiving and storing the patient Edp data sets and a calculation (sub-)server 4.2 for computer processing the patient Edp data sets stored on the data storage server 4.1.

[0025] The processing of Edp patient data sets is carried out by calculation, in particular using the calculation server 4.2, of several predefined indicators including: - for Boolean values: the existence of at least one event of each type; a number of events of each type; and / or a frequency of events of each type, for example any act performed on the patient during visits, i.e. at each visit or not, - for numerical values: a calculation of the average value of measurements taken, for example weight measurements; a calculation of the last known value of the measurement(s) taken; a calculation of trend (i.e. increase, decrease or stability) over a given period of time, typically over several months, for example over 6 months; and / or an imputation of any missing value when the value has not been measured, using the value that would be measured in a healthy patient (for example: 22 for body mass index, 0 for Conley score, etc.).

[0026] To do this, the calculation server 4.2 implements one or more data processing algorithms making it possible to process the Edp patient data sets provided by the data storage server 4.1, by proceeding as follows:

[0027] The calculation server 4.2 aggregates the information collected on the patient, during several separate medical visits, and constructs a representation of the patient's condition in the form of a collection of descriptors. These descriptors are used in conjunction with conventional classifiers (i.e. logistic regression, random forest, neural network, etc.) to categorize each patient according to their similarity to patients who have or have not fallen in the past.

[0028] In the case of logistic regression, the risk score can be defined as the weighted sum of the different patient descriptors according to the following function: ---— 1

[0029] where: • w is the vector of weighting coefficients. • y_i is the fall risk score of patient i. • X_i,t is the collection of descriptors representing the state of patient i at a instant t.

[0030] The vector w is calculated beforehand on a historical database to maximize the quality of future predictions.

[0031] In other words, the server 4 is configured to store, in particular in the data storage server 4.1, the received Edp patient data sets and to process, in particular in the calculation server 4.2, the Edp patient data sets by performing a calculation of several predefined indicators from the stored Edp patient data sets, each indicator reflecting a history of one of the events linked to the patient having occurred during a fixed elapsed period of time dt.

[0032] In addition, the calculation server 4.2 also determines (at 7) a fall risk score from the calculated indicators. The calculated score is the value y_i defined in the previous paragraph.

[0033] Display means 5, such as a computer screen or the like, cooperating with the remote server 4, in particular with the calculation server 4.2, make it possible to display fall risk information relating to the fall risk score determined for the patient P considered or for several patients P for whom respective fall risk scores have been determined, as explained above.

[0034] This then makes it possible to take preventive measures (in 8) such as, for example, sending a healthcare worker to the patient's home.

[0035] For example, [Fig.2] schematizes a graphical representation of the fall risk scores determined for several patients using a system according to the invention, such as the system of [Fig.l].

[0036] As can be seen, in the proposed embodiment, the patients P are identified by a patient identifier IDp such as their first and last name, a code (IDp#l, IDp#2, ...., IDp#n) or other.

[0037] The display means 5 are configured to simultaneously display a graphical representation of the risk of falling for the different patients concerned, namely here a bar graph 5.1, each bar 5.1 of the graph corresponding to the risk of falling for each patient P.

[0038] On the graphical representation, the patient identifications IDp, for example their name or a code, are displayed opposite the bars 5.1 of the graph.

[0039] The patients are classified here in descending order, that is to say from the patient with the highest fall risk score (at the top of the graph) to the one with the lowest fall risk score (at the bottom of the graph), that is to say that we operate and display a classification of the patients according to their fall risk score.

[0040] The length of each bar 5.1 is proportional to the fall risk score, i.e. the longer the length of a bar 5.1, the higher the score, and vice versa.

[0041] Preferably, the graphical representation comprises different colors depending on the determined fall risk score, i.e. depending on the severity, for example a green color for a low risk score and a red color for a high risk score, or even an orange color for an intermediate risk score.

[0042] Then, it is possible to select either patient to obtain additional details in addition to their fall risk score.

[0043] Of course, the graphical representation of the fall risk score of each patient P is not limited to a bar graph and can take other forms.

[0044] Furthermore, according to the invention, the system 1 can also display, for a given patient P, a graphic representation of the evolution of the fall risk score of this patient over a given period of time, for example several weeks or several months, as shown diagrammatically in [Fig.3].

[0045] To do this, the fall risk scores of each of the patients are stored over time, preferably in a computer memory of the server 4, for example in the data storage server 4.1, in the calculation server 4.2 or another location.

[0046] Being able to obtain a graphic representation of the evolution of a patient P's fall risk score over time makes it possible to determine a deterioration in their state of health and to put in place preventive actions, such as sending a nursing staff member to their home.

[0047] [Fig.4] schematizes an example of graphic representation of the factors which have led to an increase or decrease in the risk of falling for a particular patient, via a system according to the invention, such as the system of [Fig.l].

Claims

Claims

1. System (1) for determining a risk of falling in a patient (P) comprising: - means for measuring the patient's weight (2) to determine the patient's weight during visits by a healthcare professional taking place on several consecutive or non-consecutive days, - means for collecting and remotely transmitting patient data (3) to transmit, during each visit, at least one set of patient data (Edp) comprising the measured weight (Po), a patient identification (IDp), a time identifier (IDt) and at least one or more events (Ev) linked to the patient, and - at least one remote server (4) comprising reception means configured to receive the patient data sets (Edp), in which said at least one server (4) is configured to: i. store (4.1) the received patient data sets (PDS) and ii. process (4.2) patient data sets (PDS) by operating: a. a calculation of several predefined indicators from the stored patient data sets (PDS), each indicator reflecting a history of one of the patient-related events occurring during a fixed elapsed time period (dt), and b. a determination of a fall risk score from the calculated indicators, and - display means (5) cooperating with said at least one remote server (4) to display fall risk information relating to the fall risk score determined for the patient in question.

2. System according to claim 1, characterized in that said at least one remote server (4) comprises a data storage server (4.1) and a calculation server (4.2).

3. System according to claim 1, characterized in that the means for collecting and remotely transmitting patient data (3) comprise a portable device such as a laptop, digital tablet or multifunction telephone (i.e. smartphone).

4. System according to one of claims 1 or 3, characterized in that the means for collecting and remotely transmitting patient data (3) are configured to transmit said at least one set of patient data via a telecommunications protocol (6) of the GSM (4G / 5G), Sigfox, Lora or similar type.

5. System according to claim 1, characterized in that the display means (5) are configured to simultaneously display a graphical representation of the risk of falling for several patients (P).

6. System according to claim 5, characterized in that the display means (5) are configured to display a bar graph (5.1) as a graphical representation of the fall risk, each bar of the graph (5.1) corresponding to the fall risk score of each patient.

7. System according to claim 6, characterized in that the display means (5) are configured to display a ranking of the patients (P) according to their fall risk score.

8. System according to claim 1, characterized in that the display means (5) are further configured to display, for a given patient P, a graphical representation of the evolution of the fall risk score of this patient over a given period of time, for example several weeks or several months.

9. System according to one of claims 1 or 8, characterized in that the display means (5) are configured to display fall risk information comprising a numerical fall risk value, in particular a percentage, and possibly one or more associated causes.

Images