CN114937486B - IDH prediction and intervention measure recommendation multi-task model construction method and application - Google Patents
IDH prediction and intervention measure recommendation multi-task model construction method and application Download PDFInfo
- Publication number
- CN114937486B CN114937486B CN202210711223.3A CN202210711223A CN114937486B CN 114937486 B CN114937486 B CN 114937486B CN 202210711223 A CN202210711223 A CN 202210711223A CN 114937486 B CN114937486 B CN 114937486B
- Authority
- CN
- China
- Prior art keywords
- dialysis
- idh
- data
- task
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000010276 construction Methods 0.000 title abstract description 3
- 210000002569 neuron Anatomy 0.000 claims abstract description 89
- 238000001631 haemodialysis Methods 0.000 claims abstract description 81
- 230000000322 hemodialysis Effects 0.000 claims abstract description 81
- 238000012360 testing method Methods 0.000 claims abstract description 64
- 208000001953 Hypotension Diseases 0.000 claims abstract description 62
- 230000036543 hypotension Effects 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000011282 treatment Methods 0.000 claims abstract description 31
- 230000008569 process Effects 0.000 claims abstract description 23
- 238000000502 dialysis Methods 0.000 claims description 159
- 210000004369 blood Anatomy 0.000 claims description 58
- 239000008280 blood Substances 0.000 claims description 58
- 238000000108 ultra-filtration Methods 0.000 claims description 30
- 230000006870 function Effects 0.000 claims description 28
- 210000002700 urine Anatomy 0.000 claims description 28
- 238000011038 discontinuous diafiltration by volume reduction Methods 0.000 claims description 25
- 238000001802 infusion Methods 0.000 claims description 25
- 230000003827 upregulation Effects 0.000 claims description 25
- 239000002504 physiological saline solution Substances 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 19
- 230000035488 systolic blood pressure Effects 0.000 claims description 18
- 238000003491 array Methods 0.000 claims description 15
- 239000000725 suspension Substances 0.000 claims description 15
- DDRJAANPRJIHGJ-UHFFFAOYSA-N creatinine Chemical compound CN1CC(=O)NC1=N DDRJAANPRJIHGJ-UHFFFAOYSA-N 0.000 claims description 14
- 238000003745 diagnosis Methods 0.000 claims description 13
- 238000012423 maintenance Methods 0.000 claims description 12
- 230000017531 blood circulation Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 10
- 102000009027 Albumins Human genes 0.000 claims description 8
- 108010088751 Albumins Proteins 0.000 claims description 8
- BHPQYMZQTOCNFJ-UHFFFAOYSA-N Calcium cation Chemical compound [Ca+2] BHPQYMZQTOCNFJ-UHFFFAOYSA-N 0.000 claims description 8
- NPYPAHLBTDXSSS-UHFFFAOYSA-N Potassium ion Chemical compound [K+] NPYPAHLBTDXSSS-UHFFFAOYSA-N 0.000 claims description 8
- FKNQFGJONOIPTF-UHFFFAOYSA-N Sodium cation Chemical compound [Na+] FKNQFGJONOIPTF-UHFFFAOYSA-N 0.000 claims description 8
- 229910001424 calcium ion Inorganic materials 0.000 claims description 8
- HVYWMOMLDIMFJA-DPAQBDIFSA-N cholesterol Chemical compound C1C=C2C[C@@H](O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2 HVYWMOMLDIMFJA-DPAQBDIFSA-N 0.000 claims description 8
- 230000035487 diastolic blood pressure Effects 0.000 claims description 8
- 229910001414 potassium ion Inorganic materials 0.000 claims description 8
- 230000036387 respiratory rate Effects 0.000 claims description 8
- 229910001415 sodium ion Inorganic materials 0.000 claims description 8
- 229940109239 creatinine Drugs 0.000 claims description 7
- 210000003743 erythrocyte Anatomy 0.000 claims description 7
- 230000037396 body weight Effects 0.000 claims description 6
- 102000001554 Hemoglobins Human genes 0.000 claims description 5
- 108010054147 Hemoglobins Proteins 0.000 claims description 5
- 206010020772 Hypertension Diseases 0.000 claims description 5
- 230000004872 arterial blood pressure Effects 0.000 claims description 5
- 206010012601 diabetes mellitus Diseases 0.000 claims description 5
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 claims description 4
- ZAMOUSCENKQFHK-UHFFFAOYSA-N Chlorine atom Chemical compound [Cl] ZAMOUSCENKQFHK-UHFFFAOYSA-N 0.000 claims description 4
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 claims description 4
- DGAQECJNVWCQMB-PUAWFVPOSA-M Ilexoside XXIX Chemical compound C[C@@H]1CC[C@@]2(CC[C@@]3(C(=CC[C@H]4[C@]3(CC[C@@H]5[C@@]4(CC[C@@H](C5(C)C)OS(=O)(=O)[O-])C)C)[C@@H]2[C@]1(C)O)C)C(=O)O[C@H]6[C@@H]([C@H]([C@@H]([C@H](O6)CO)O)O)O.[Na+] DGAQECJNVWCQMB-PUAWFVPOSA-M 0.000 claims description 4
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 claims description 4
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 claims description 4
- LEHOTFFKMJEONL-UHFFFAOYSA-N Uric Acid Chemical compound N1C(=O)NC(=O)C2=C1NC(=O)N2 LEHOTFFKMJEONL-UHFFFAOYSA-N 0.000 claims description 4
- TVWHNULVHGKJHS-UHFFFAOYSA-N Uric acid Natural products N1C(=O)NC(=O)C2NC(=O)NC21 TVWHNULVHGKJHS-UHFFFAOYSA-N 0.000 claims description 4
- PNNCWTXUWKENPE-UHFFFAOYSA-N [N].NC(N)=O Chemical compound [N].NC(N)=O PNNCWTXUWKENPE-UHFFFAOYSA-N 0.000 claims description 4
- 239000003146 anticoagulant agent Substances 0.000 claims description 4
- 229940127219 anticoagulant drug Drugs 0.000 claims description 4
- 238000004820 blood count Methods 0.000 claims description 4
- 230000036760 body temperature Effects 0.000 claims description 4
- 239000011575 calcium Substances 0.000 claims description 4
- 229910052791 calcium Inorganic materials 0.000 claims description 4
- 239000000460 chlorine Substances 0.000 claims description 4
- 229910052801 chlorine Inorganic materials 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 4
- 239000008103 glucose Substances 0.000 claims description 4
- 210000000265 leukocyte Anatomy 0.000 claims description 4
- 208000012866 low blood pressure Diseases 0.000 claims description 4
- 239000011574 phosphorus Substances 0.000 claims description 4
- 229910052698 phosphorus Inorganic materials 0.000 claims description 4
- 239000011591 potassium Substances 0.000 claims description 4
- 229910052700 potassium Inorganic materials 0.000 claims description 4
- 102000004169 proteins and genes Human genes 0.000 claims description 4
- 108090000623 proteins and genes Proteins 0.000 claims description 4
- 239000011734 sodium Substances 0.000 claims description 4
- 229910052708 sodium Inorganic materials 0.000 claims description 4
- 150000003626 triacylglycerols Chemical class 0.000 claims description 4
- 229940116269 uric acid Drugs 0.000 claims description 4
- 230000004584 weight gain Effects 0.000 claims description 4
- 235000019786 weight gain Nutrition 0.000 claims description 4
- 235000012000 cholesterol Nutrition 0.000 claims description 3
- 238000000748 compression moulding Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 208000021822 hypotensive Diseases 0.000 claims 11
- 230000001077 hypotensive effect Effects 0.000 claims 11
- 208000020372 Infective dermatitis associated with HTLV-1 Diseases 0.000 abstract description 158
- 238000004393 prognosis Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000001154 acute effect Effects 0.000 description 5
- 208000019901 Anxiety disease Diseases 0.000 description 4
- 206010008479 Chest Pain Diseases 0.000 description 4
- 206010009866 Cold sweat Diseases 0.000 description 4
- 206010013954 Dysphoria Diseases 0.000 description 4
- 208000000059 Dyspnea Diseases 0.000 description 4
- 206010013975 Dyspnoeas Diseases 0.000 description 4
- 208000007101 Muscle Cramp Diseases 0.000 description 4
- 206010028813 Nausea Diseases 0.000 description 4
- 241001282135 Poromitra oscitans Species 0.000 description 4
- 208000005392 Spasm Diseases 0.000 description 4
- 206010047700 Vomiting Diseases 0.000 description 4
- 206010048232 Yawning Diseases 0.000 description 4
- 206010000059 abdominal discomfort Diseases 0.000 description 4
- 230000036506 anxiety Effects 0.000 description 4
- 230000007211 cardiovascular event Effects 0.000 description 4
- 208000002173 dizziness Diseases 0.000 description 4
- 230000008693 nausea Effects 0.000 description 4
- 230000001052 transient effect Effects 0.000 description 4
- 230000008673 vomiting Effects 0.000 description 4
- 208000003443 Unconsciousness Diseases 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 210000004379 membrane Anatomy 0.000 description 3
- 239000012528 membrane Substances 0.000 description 3
- 208000037157 Azotemia Diseases 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 239000003792 electrolyte Substances 0.000 description 2
- 239000012510 hollow fiber Substances 0.000 description 2
- 230000002503 metabolic effect Effects 0.000 description 2
- 208000009852 uremia Diseases 0.000 description 2
- 208000009304 Acute Kidney Injury Diseases 0.000 description 1
- 206010018367 Glomerulonephritis chronic Diseases 0.000 description 1
- 208000033626 Renal failure acute Diseases 0.000 description 1
- FAPWRFPIFSIZLT-UHFFFAOYSA-M Sodium chloride Chemical compound [Na+].[Cl-] FAPWRFPIFSIZLT-UHFFFAOYSA-M 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 201000011040 acute kidney failure Diseases 0.000 description 1
- 208000012998 acute renal failure Diseases 0.000 description 1
- 239000003513 alkali Substances 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 208000020832 chronic kidney disease Diseases 0.000 description 1
- 208000022831 chronic renal failure syndrome Diseases 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000385 dialysis solution Substances 0.000 description 1
- 239000008151 electrolyte solution Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 210000003200 peritoneal cavity Anatomy 0.000 description 1
- 210000002381 plasma Anatomy 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 238000010837 poor prognosis Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012959 renal replacement therapy Methods 0.000 description 1
- 239000011780 sodium chloride Substances 0.000 description 1
- 238000001179 sorption measurement Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000003440 toxic substance Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/40—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention relates to a construction method and application of an IDH prediction and intervention measure recommendation multitasking model, belonging to the field of artificial intelligence in the medical industry, comprising the following steps: s1, collecting time-invariant data before hemodialysis and forming input data by the time-invariant data; s2, setting a label for each piece of input data, wherein each label corresponds to a learning task; s3, constructing a centrally-penetrating hypotension prediction and intervention measure recommendation multi-task model, wherein the model comprises an auxiliary task X, an auxiliary task Y and a main task Z; the model framework of each task is equally divided into an input layer, a hidden layer and an output layer, and each layer is composed of a plurality of neurons; s4, constructing training set data and test set data to train the model; s5, verifying a multitasking model; the risk of IDH occurrence of patients is reduced by means of early prediction and early intervention, so that the aims of preventing and treating IDH in the MHD treatment process and improving the prognosis of MHD patients are fulfilled.
Description
Technical Field
The invention relates to a method for constructing a medical prediction model, belongs to the field of artificial intelligence in the medical industry, and in particular relates to a method for constructing a model for predicting and recommending intervention measures in maintainance hemodialysis through by adopting a multi-task learning method and application of the model.
Background
The maintenance hemodialysis treatment is a transitional method for saving the life of a patient by hemodialysis or peritoneal dialysis and prolonging the life of a uremic patient. Patients undergoing maintenance hemodialysis include uremia caused by the development of chronic nephritis, and other uremia caused by diabetes and hypertension are common causes of the patient undergoing maintenance hemodialysis. The maintenance dialysis treatment is divided into hemodialysis and peritoneal dialysis:
hemodialysis (HD) is one of the renal replacement therapies for patients with acute and chronic renal failure. The method comprises the steps of draining in-vivo blood to the outside of the body, passing through a dialyzer consisting of innumerable hollow fibers, wherein the blood and electrolyte solution (dialysate) with similar concentration of the body are in and out of the hollow fibers, and carrying out substance exchange by the dispersion, ultrafiltration, adsorption and convection principles to remove metabolic wastes in the body and maintain the balance of the electrolyte and acid and alkali; simultaneously, the excessive moisture in the body is removed, and the purified blood is returned to the body.
Peritoneal Dialysis (PD) is to regularly and regularly fill a prepared dialysate into a peritoneal cavity of a patient through a catheter by utilizing the characteristic that a peritoneal membrane is used as a semi-permeable membrane under the action of gravity, and the solute on the high concentration side moves to the low concentration side (dispersion effect) due to the concentration gradient difference of the solute on the two sides of the peritoneal membrane; moisture moves from the hypotonic side to the hypertonic side (osmosis). The peritoneal dialysis solution is continuously replaced to achieve the purposes of removing metabolic products and toxic substances in the body and correcting water and electrolyte balance disturbance.
Hypotension (IDH) in dialysis is one of the most common complications during Maintenance Hemodialysis (MHD) treatment, and in dialysis, when the patient has a systolic blood pressure drop of 20mmHg or an average arterial blood pressure drop of 10mmHg, the patient can suffer dizziness, dysphoria, anxiety, pale complexion, yawning, nausea, vomiting, chest distress, increased heart rate, abdominal discomfort, cold sweat, severe patients can have dyspnea, blacking, muscle spasms, even loss of transient consciousness, and acute cardiovascular events can be caused in severe cases, increasing the risk of death. Studies show that the occurrence rate of IDH in MHD treatment patients in China is about 39%, and the frequent occurrence rate of IDH is an important factor of poor prognosis of MHD patients, so that the prevention and treatment of IDH in the MHD treatment process are very important.
Disclosure of Invention
In order to solve the technical problems, the invention provides a method for constructing a maintenance hemodialysis hypotension prediction and intervention recommendation multitasking model, which comprises the following steps:
s1, collecting input data; collecting time-invariant data before hemodialysis and time-varying data at each time during hemodialysis as input data; forming an input data by each time-invariant data and each time-variant data together;
S2, setting a label for each piece of input data; setting 3 labels for each piece of input data, namely a label A, a label B and a label C; each label corresponds to a learning task;
s3, constructing a centrally-penetrating hypotension prediction and intervention measure recommendation multi-task model; the multi-task model comprises an auxiliary task X, an auxiliary task Y and a main task Z; the model framework of each task is equally divided into an input layer, a hidden layer and an output layer, and each layer is composed of a plurality of neurons; each neuron in the hidden layer and the output layer has a weight, and the weight is obtained in the process of model training; each neuron of the output layer can have different weight calculations on the data output by the hidden layer, and the final result is output, wherein the number of the neurons of the output layer is determined by the number of the label categories;
s4, training a model; constructing training set data and test set data from the input data collected in the step S1, constructing a loss function, inputting the training set data into a model, and training the model;
s5, verifying a multitasking model; the multi-task model calculates the test set data obtained in the step S4 to obtain the results of three tasks of the test set data, calculates the accuracy rate, recall rate and precision rate of each task respectively, evaluates the effect of the multi-task model through the three indexes, and obtains the qualified IDH prediction and intervention recommended multi-task model when the accuracy rate, recall rate and precision rate of each task reach preset values.
Further, the method comprises the steps of,
the unchanged data in step S1 includes: measured sign data, personal information, medical history records, test results, historical dialysis records, current hemodialysis orders, prior to patient hemodialysis; wherein,
measuring the sign data includes: anterior body weight, anterior heart rate, anterior dry body weight, anterior pulse, anterior systolic pressure, anterior diastolic pressure, and anterior respiratory rate;
the personal information includes: age, sex, height, dialysis age;
the medical history record includes: a history of hypertension and diabetes;
the checking result includes: blood white count, hemoglobin, red blood cell count, blood glucose, blood albumin, blood total cholesterol, blood triglycerides, blood creatinine, blood uric acid, blood urea nitrogen, blood potassium, blood sodium, blood calcium, blood phosphorus, blood chlorine, urine white blood cells, urine proteins, urine red blood cells, urine creatinine, urine occult blood, urine microalbumin, urine albumin;
the history dialysis record includes: weight gain during dialysis, last dialysis time, last machine weight, number of times IDH occurs in seven days and number of times IDH occurs in thirty days;
the current hemodialysis prescription includes: dialysis mode, anticoagulant dose, ultrafiltration volume, dialysis duration, dialysate potassium ion concentration, dialysate calcium ion concentration, dialysate sodium ion concentration, dialysate conductivity, blood flow;
The time-varying data includes: during hemodialysis of a patient, monitoring physical signs, dialysis treatment parameters and dialysis machine parameters at fixed time intervals; wherein,
the interval monitoring sign data includes: current body temperature, current heart rate, current systolic pressure, current diastolic pressure, current pulse, current respiratory rate;
dialysis treatment parameters included: current ultrafiltered, current dialysis duration, current dialysate potassium ion concentration, current dialysate calcium ion concentration, current dialysate sodium ion concentration, current dialysate conductivity, current blood flow;
the dialysis machine parameters included: current arterial pressure, current venous pressure, current cross-compression molding, current blood flow.
Further, step S2 includes the following sub-steps:
s21, setting diagnosis standard of IDH:
(1) an IDH compliant intervention, and a systolic pressure drop of more than 20mmHg from the pre-permeabilized systolic pressure;
(2) no intervention, but a systolic blood pressure of less than 90mmHg;
s22, setting a label A, a label B and a label C for each piece of input data according to the diagnosis standard of IDH;
the label A is: whether IDH occurs in the dialysis;
the label B is: whether IDH occurs at the next moment;
the label C is: an intervention measure of IDH at the next moment; the IDH intervention method comprises the following steps: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, saline infusion, up-regulation of conductivity.
Further, step S22 includes the following sub-steps:
s221, collecting IDH intervention data in dialysis, wherein the IDH intervention data is an intervention measure of the input data in the dialysis process; the intervention measures of the IDH intervention data are also as follows: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity;
s222, setting a label A; judging whether the input data of all times of the dialysis is IDH or not at each time according to the diagnosis standard of IDH;
s223, setting a label B; judging whether the input data of all times of the dialysis is IDH or not at each time according to the diagnosis standard of IDH; the label B at the previous moment of the moment of judging the IDH is marked as the next moment of generating the IDH, and the labels B at the other moments mark the next moment of not generating the IDH;
s224, setting a label C; the input data acquisition tag C is set according to the intervention data of the IDH.
Further, in step S3, it is predicted whether IDH occurs in the present dialysis as an auxiliary task X; predicting the intervention measure of IDH at the next moment as an auxiliary task Y; predicting whether IDH occurs at the next moment as a main task Z;
the model framework of the auxiliary task X is divided into an input layer X1, a hidden layer X2 and an output layer X3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer X2 and the output layer X3 has a weight which is obtained in the model training process;
1. The auxiliary task X model comprises the following structure:
(1) Input layer: the data input by the input layer X1 of the auxiliary task X are time-invariant data before hemodialysis, and the number of neurons of the input layer X1 is time-invariant data before hemodialysis;
(2) Hidden layer: each neuron of the hidden layer X2 may have different weight calculations on the data input by the input layer X1, so as to further favor the prediction of a certain task tag; in the invention, the weight value of the hidden layer X2 of the auxiliary task X is biased to predict whether IDH occurs in the dialysis;
setting:
the output array of the input layer X1 of the auxiliary task X is X1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer X2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer X2 be X2 i Then:
X2 i ＝X1 i *W i ；
(3) Output layer: one neuron of the output layer X3 is used for generating no IDH in the present dialysis, and 0 is used for generating IDH in the present dialysis; the neuron of the output layer X3 has different weight calculations on the data output by the hidden layer, and outputs a final result;
setting:
the neuron of the output layer X3 comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then the first time period of the first time period,
the model of the auxiliary task Y includes the following structure:
the model framework of the auxiliary task Y is divided into an input layer Y1, a hidden layer Y2 and an output layer Y3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Y2 and the output layer Y3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer Y1 of the auxiliary task Y is the same as the data input by the input layer of the main task Z1, and is time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Y1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period;
(2) Hidden layer: each neuron of the hidden layer Y2 may have different weight calculation on the data input by the input layer Y1, so as to further favor the prediction of a certain task label;
setting:
the output array of the input layer Y1 of the auxiliary task Y is Y1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer Y2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer Y2 be Y2 i Then:
Y2 i ＝Y1 i *W i ；
in the invention, the weight value of the hidden layer Y2 of the auxiliary task Y is biased to predict the intervention measure of IDH at the next moment;
(3) Output layer: five output layer Y3 neurons respectively correspond to intervention measures of IDH at the next moment such as no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity and the like;
the neuron of the output layer Y3 has different weight calculations on the data output by the hidden layer and outputs a final result;
setting:
the neuron of the output layer Y3 comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then the first time period of the first time period,
3. the main task Z model comprises the following structure:
the model framework of the main task Z is divided into an input layer Z1, a hidden layer Z2 and an output layer Z3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Z2 and the output layer Z3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer of the main task Z1 is the same as the data input by the input layer Y1 of the auxiliary task Y, and is time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Z1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period;
(2) Hidden layer: each neuron of the hidden layer Z2 may have different weight calculation on the data input by the input layer Z1, so that prediction of a certain task label is favored, in the present invention, the weight value of the hidden layer Z2 of the main task Z is favored to predict whether IDH occurs at the next moment;
(3) Output layer: one output layer Z3 neuron of the main task Z is as follows: no IDH occurs at the next time, denoted by "0", or IDH occurs at the next time, denoted by "1";
taking an output array of a hidden layer X2 of an auxiliary task X, an output array of a hidden layer Y2 of an auxiliary task Y and an output array of a hidden layer Z2 of a main task Z as inputs of an output layer Z3 of the main task Z;
setting:
the output array of the hidden layer X2 of the auxiliary task X is X2 i I is the number of numerical values in the output array, and i is 1-n;
output array Y2 of hidden layer Y2 of auxiliary task Y i I is the number of numerical values in the output array, and i is 1-n;
the output array of the hidden layer Z2 of the main task Z is Z2 i I is the number of numerical values in the output array, and i is 1-n;
the number of the output array of the hidden layer X2 of the auxiliary task X, the number of the output array of the hidden layer Y2 of the auxiliary task Y and the number of the numerical values in the output array of the hidden layer Z2 of the main task Z are the same, i is 1-n;
The neuron of the main task Z output layer Z3 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
setting the output array as X2 i Output array Y2 i Output array Z2 i The same in (a)The values in the sequence are added to obtain an input array Q of the output layer Z3 of the main task Z i ；
Further, step S4 includes the following sub-steps:
s41, constructing training set data and test set data;
sequencing the input data collected in the step S1 according to dialysis time, taking the first 80% as training set data and the last 20% as test set data;
s42, constructing a loss function;
firstly, constructing a loss function by adopting cross entropy for an auxiliary task X, an auxiliary task Y and a main task Z;
then constructing a model total loss function, wherein the model total loss function is a weighted sum of the loss functions of the auxiliary task X, the auxiliary task Y and the main task Z; wherein,
the weight ratio of the cross entropy loss function of the main task Z to the auxiliary tasks X and Y is 2:1:1, a step of;
s43, calculating weight parameters of corresponding neurons when the loss function takes the minimum value by adopting a gradient descent method for training set data;
The neuron is a neuron in the model constructed in S3.
Further, in step S5:
the accuracy of the auxiliary task A is the ratio of the number of the model correctly predicted IDH occurrence and the IDH non-occurrence in the test set data and the total number of the test data;
the recall rate of the auxiliary task A is the ratio of the number of IDH occurrence correctly predicted by the model in the test set data to the total number of IDH occurrence in the test set data;
the accuracy of the auxiliary task A is the ratio of the number of IDH occurrence correctly predicted by the model in the test set data to the total number of IDH occurrence predicted by the model;
the accuracy of the auxiliary task B is the ratio of the sum of the numbers of the model correct prediction without intervention, suspension ultrafiltration, volume reduction, physiological saline infusion and up-regulation conductivity in the test set data to the total number of the test data;
the recall rate of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data respectively;
the accuracy of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the model predicted non-intervention, the total number of the suspended ultrafiltration, the total number of the volume reduction, the total number of the physiological saline infusion and the total number of the up-regulation conductivity;
The accuracy of the main task C is the ratio of the number of the occurrence of IDH at the next moment and the non-occurrence of IDH at the next moment and the total quantity of the test data, which are correctly predicted by the model in the test set data;
the recall rate of the main task C is the ratio of the number of IDH occurrence at the next moment in the test set data to the total number of IDH occurrence at the next moment in the test set data, which is correctly predicted by the model in the test set data;
the accuracy of the primary task C is the ratio of the number of IDH occurrences in the test set data that the model correctly predicts for the next time to the total number of IDH occurrences that the model predicts for the next time.
Furthermore, the invention also provides an application of the IDH prediction and intervention recommended multitasking model constructed by the method in predicting IDH in maintenance hemodialysis treatment.
The beneficial effects are that: the invention discloses a method for recommending a multitask model based on multitask learning through midrange hypotension prediction and intervention measures, which reduces the risk of IDH occurrence of a patient in a mode of early prediction and early intervention, thereby achieving the purposes of preventing and treating IDH in the MHD treatment process and improving the prognosis of the MHD patient, and specifically:
in one aspect, it is not only possible to predict whether IDH is occurring in the present dialysis prior to patient hemodialysis, to increase the medical care's attention to patients who are likely to develop IDH.
On the other hand, the model can also be used for predicting whether IDH occurs at the next monitoring moment in the dialysis process according to the monitored sign and dialysis treatment parameters during hemodialysis.
Thirdly, by adopting the model of the invention, a high-quality IDH intervention scheme can be provided for medical staff, and a high-quality treatment scheme recommendation, namely a high-quality IDH intervention measure recommendation, is provided when IDH occurs in hemodialysis, so that the problems that the medical staff selects an improper treatment scheme or can not give the treatment scheme in a short time or the medical staff without experience does not know how to treat the patient, such as dizziness, dysphoria, anxiety, pale complexion, yawning, nausea, vomiting, chest distress, heart rate increase, abdominal discomfort, cold sweat, dyspnea, blackness, muscle spasm and even transient consciousness loss are solved, and acute cardiovascular events, even death and the like, are caused when serious patients are serious.
Fourth, the invention fuses the output of the hidden layer of the auxiliary task and the output of the hidden layer of the main task, and the fusion result is used as the input of the output layer of the main task, thereby realizing the weight sharing of the main task and the auxiliary task.
Fifth, the model of the invention can simultaneously realize the prediction of a plurality of tasks, thereby saving the storage cost and maintenance cost of the model, and the time cost and calculation cost of model prediction.
Drawings
FIG. 1 is a schematic diagram of a multitasking model for predicting and recommending intervention in the middle-low blood pressure.
Detailed Description
Example 1: the invention provides a method for constructing a maintenance blood plasma hypotension prediction and intervention recommended multitasking model and application thereof, comprising the following steps:
s1, collecting input data; collecting time-invariant data before hemodialysis and time-varying data at each time during hemodialysis as input data; forming an input data by each time-invariant data and each time-variant data together;
at least 1 piece of input data;
(1) The time invariant data includes: measured sign data, personal information, medical history records, test results, historical dialysis records, current hemodialysis orders, prior to patient hemodialysis;
wherein ,
measuring the sign data includes: anterior body weight, anterior heart rate, anterior dry body weight, anterior pulse, anterior systolic pressure, anterior diastolic pressure, and anterior respiratory rate;
the personal information includes: age, sex, height, dialysis age;
The medical history record includes: a history of hypertension and diabetes;
the checking result includes: blood white count, hemoglobin, red blood cell count, blood glucose, blood albumin, blood total cholesterol, blood triglycerides, blood creatinine, blood uric acid, blood urea nitrogen, blood potassium, blood sodium, blood calcium, blood phosphorus, blood chlorine, urine white blood cells, urine proteins, urine red blood cells, urine creatinine, urine occult blood, urine microalbumin, urine albumin;
the history dialysis record includes: weight gain during dialysis, last dialysis time, last machine weight, number of times IDH occurs in seven days and number of times IDH occurs in thirty days;
the current hemodialysis prescription includes: dialysis mode, anticoagulant dose, ultrafiltration volume, dialysis duration, dialysate potassium ion concentration, dialysate calcium ion concentration, dialysate sodium ion concentration, dialysate conductivity, blood flow.
(2) The time-varying data includes: during hemodialysis of a patient, monitoring physical signs, dialysis treatment parameters and dialysis machine parameters at fixed time intervals; wherein,
the interval monitoring sign data includes: current body temperature, current heart rate, current systolic pressure, current diastolic pressure, current pulse, current respiratory rate;
Dialysis treatment parameters included: current ultrafiltered, current dialysis duration, current dialysate potassium ion concentration, current dialysate calcium ion concentration, current dialysate sodium ion concentration, current dialysate conductivity, current blood flow.
The dialysis machine parameters included: current arterial pressure, current venous pressure, current cross-compression molding, current blood flow.
S2, setting a label for each piece of input data; setting 3 labels for each piece of input data, namely a label A, a label B and a label C; each label corresponds to a learning task;
the label A is: whether IDH occurs in the dialysis;
the label B is: whether IDH occurs at the next moment;
the label C is: an intervention measure of IDH at the next moment; the IDH intervention method comprises the following steps: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity;
s21, setting diagnosis standard of IDH:
(1) an IDH compliant intervention, and a systolic pressure drop of more than 20mmHg from the pre-permeabilized systolic pressure;
(2) no intervention, but a systolic blood pressure of less than 90mmHg;
s22, setting a label A, a label B and a label C for each piece of input data according to the diagnosis standard of IDH;
specifically, the method comprises the following substeps:
S221, collecting IDH intervention data in dialysis, wherein the IDH intervention data is an intervention measure of the input data in the dialysis process; the intervention measures of the IDH intervention data are also as follows: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity;
each piece of input data corresponds to one piece of IDH intervention data; the input data, IDH intervention data, a label A, a label B and a label C form a complete data;
s222, setting a label A; judging whether the input data of all times of the dialysis is IDH or not at each time according to the diagnosis standard of IDH;
the input data of all times of the dialysis comprises N pieces of input data, wherein N is at least 1;
specifically, each moment of the present dialysis corresponds to one piece of input data, and N moments of the present dialysis correspond to N pieces of input data; the time is a time point of a fixed time interval, and the fixed time interval comprises time values of hours, minutes, seconds, milliseconds and the like;
if IDH occurs at one moment, marking the labels A of the input data at all moments of the dialysis as IDH, namely marking the labels A of the N input data of the dialysis as IDH;
if IDH does not occur at each moment, marking the label A of the input data at all moments of the dialysis as IDH does not occur; namely, marking the labels A of N pieces of input data of the dialysis as non-IDH;
Further, tag a is represented by 0 or 1, 0 being no IDH occurrence, 1 being IDH occurrence; the label B is represented by 0 or 1, 0 being that IDH does not occur at the next time, and 1 being that IDH occurs at the next time.
S223, setting a label B; judging whether the input data of all times of the dialysis is IDH or not at each time according to the diagnosis standard of IDH; the label B at the previous moment of the moment of judging the IDH is marked as the next moment of generating the IDH, and the labels B at the other moments mark the next moment of not generating the IDH;
similarly, the input data at all times of the dialysis at this time comprises N pieces of input data, wherein N is at least 1;
specifically, each moment of the present dialysis corresponds to one piece of input data, and N moments of the present dialysis correspond to N pieces of input data; the time is a time point of a fixed time interval, and the fixed time interval comprises hours, minutes, seconds, milliseconds or the like;
for example:
example 1: taking n=5 as an example, i.e., the number of pieces of input data is set to 5; if the input data with IDH is n=4 according to the diagnostic criteria of IDH, label B in the input data with n=3 is labeled as: IDH occurs at the next time; the label B of the remaining input data is labeled: IDH does not occur at the next moment;
example 2: taking n=5 as an example, i.e., the number of pieces of input data is set to 5; setting the input data where IDH occurs as n=4, and n=5, according to the diagnostic criteria judgment of IDH;
The tag B in the input data of n=3 and the input data of n=4 is marked as: IDH occurs at the next time; the label B of the remaining input data is labeled: IDH does not occur at the next moment;
the invention judges whether the input data of all times of the dialysis is IDH or not at each time according to the diagnosis standard of IDH; the label B at the moment which is the moment of IDH is judged to be the next moment to generate IDH, and the label B at the rest moment is marked to be the next moment to not generate IDH, so that the problem that whether IDH occurs at the next moment in the dialysis process is solved, and the problem of predicting IDH risk in real time in the hemodialysis process is solved.
S224, setting a label C; setting an input data acquisition tag C according to the intervention measure data of the IDH;
further, tag C is represented by a one-hot code, using five digits to code five interventions; examples:
10000 means no intervention;
01000 indicates suspension of ultrafiltration;
00100 represents volume reduction;
00010 represents physiological saline infusion;
00001 represents the up-regulation conductance.
Further, the method comprises the steps of,
firstly, judging what kind of intervention measures belong to next input data of input data with a label B of 1 of each input data according to the intervention measures of IDH, then, marking a label C of the input data with the label B of 1 as a single-heat code of the corresponding intervention measure, and marking the labels C at the rest moments as single-heat codes without intervention.
For example:
m pieces of data are provided, m=6;
and (3) marking the label B in the 4 th input data as 1, judging what kind of intervention measure the 5 th input data belongs to according to the intervention measure of IDH, marking the label C of the 4 th input data as the single-heat coding of the corresponding intervention measure, and marking the labels C at the rest moments as the single-heat coding without intervention.
According to the invention, the input data acquisition label C is set according to IDH intervention measure data, so that a high-quality IDH intervention scheme can be provided for medical staff, and a treatment scheme is provided for the occurrence of IDH in hemodialysis, thereby solving the problems that the medical staff selects an improper treatment scheme when IDH occurs in hemodialysis, or can not give a treatment scheme in a short time, or the medical staff without experience does not know how to treat the patient, which is caused by dizziness, dysphoria, anxiety, pale complexion, yawning, nausea, vomiting, chest distress, heart rate increase, abdominal discomfort, cold sweat, and serious patients can have dyspnea, blackness, muscle spasm, even transient consciousness loss, and cause acute cardiovascular events, even death, and the like.
Meanwhile, the invention solves the problem of providing a predicted high-quality IDH intervention scheme by firstly judging which intervention measure belongs to the next input data of the input data with the label B of each input data marked as 1 according to the intervention measure of the IDH and then marking the label C of the input data with the label B marked as 1 as the corresponding intervention measure in a single-hot coding mode.
S3, constructing a centrally-penetrating hypotension prediction and intervention measure recommendation multi-task model;
the multi-task model comprises an auxiliary task X, an auxiliary task Y and a main task Z;
as shown in fig. 1, the model architecture of each task is equally divided into an input layer, a hidden layer and an output layer, and each layer is composed of a plurality of neurons; each neuron in the hidden layer and the output layer has a weight, and the weight is obtained in the process of model training; each neuron of the output layer can have different weight calculations on the data output by the hidden layer, and the final result is output, wherein the number of the neurons of the output layer is determined by the number of the label categories;
setting:
predicting whether IDH occurs in the present dialysis as an auxiliary task X;
predicting the intervention measure of IDH at the next moment as an auxiliary task Y;
predicting whether IDH occurs at the next moment as a main task Z;
then:
1. the model framework of the auxiliary task X is divided into an input layer X1, a hidden layer X2 and an output layer X3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer X2 and the output layer X3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer X1 of the auxiliary task X are time-invariant data before hemodialysis, and the number of neurons of the input layer X1 is time-invariant data before hemodialysis;
(2) Hidden layer: each neuron of the hidden layer X2 may have different weight calculations on the data input by the input layer X1, so as to further favor the prediction of a certain task tag; in the invention, the weight value of the hidden layer X2 of the auxiliary task X is biased to predict whether IDH occurs in the dialysis;
setting:
the output array of the input layer X1 of the auxiliary task X is X1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer X2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer X2 be X2 i Then:
X2 i ＝X1 i *W i ；
(3) Output layer: one neuron of the output layer X3 is used for generating no IDH in the present dialysis, and 0 is used for generating IDH in the present dialysis; the neuron of the output layer X3 has different weight calculations on the data output by the hidden layer, and outputs a final result;
setting:
the neuron of the output layer X3 comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then the first time period of the first time period,
2. the model framework of the auxiliary task Y is divided into an input layer Y1, a hidden layer Y2 and an output layer Y3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Y2 and the output layer Y3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer Y1 of the auxiliary task Y is the same as the data input by the input layer of the main task Z1, and is time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Y1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period;
(2) Hidden layer: each neuron of the hidden layer Y2 may have different weight calculation on the data input by the input layer Y1, so as to further favor the prediction of a certain task label;
setting:
the output array of the input layer Y1 of the auxiliary task Y is Y1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer Y2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer Y2 be Y2 i Then:
Y2 i ＝Y1 i *W i ；
in the invention, the weight value of the hidden layer Y2 of the auxiliary task Y is biased to predict the intervention measure of IDH at the next moment;
(3) Output layer: five output layer Y3 neurons respectively correspond to intervention measures of IDH at the next moment such as no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity and the like;
The neuron of the output layer Y3 has different weight calculations on the data output by the hidden layer and outputs a final result;
setting:
output layer Y3The neuron comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then the first time period of the first time period,
3. the model framework of the main task Z is divided into an input layer Z1, a hidden layer Z2 and an output layer Z3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Z2 and the output layer Z3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer of the main task Z1 is the same as the data input by the input layer Y1 of the auxiliary task Y, the data are time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Z1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period.
(2) Hidden layer: each neuron of the hidden layer Z2 may have different weight calculation on the data input by the input layer Z1, so that prediction of a certain task label is favored, in the present invention, the weight value of the hidden layer Z2 of the main task Z is favored to predict whether IDH occurs at the next moment;
(3) Output layer: one output layer Z3 neuron of the main task Z is as follows: no IDH occurs at the next time, denoted by "0", or IDH occurs at the next time, denoted by "1";
as shown in figure 1 of the drawings,
taking an output array of a hidden layer X2 of an auxiliary task X, an output array of a hidden layer Y2 of an auxiliary task Y and an output array of a hidden layer Z2 of a main task Z as inputs of an output layer Z3 of the main task Z;
setting:
the output array of the hidden layer X2 of the auxiliary task X is X2 i I is the number of numerical values in the output array, and i is 1-n;
output array Y of hidden layer Y2 of auxiliary task Y2 i I is the number of numerical values in the output array, and i is 1-n;
the output array of the hidden layer Z2 of the main task Z is Z2 i I is the number of numerical values in the output array, and i is 1-n;
the number of the output array of the hidden layer X2 of the auxiliary task X, the number of the output array of the hidden layer Y2 of the auxiliary task Y and the number of the numerical values in the output array of the hidden layer Z2 of the main task Z are the same, i is 1-n;
the neuron of the main task Z output layer Z3 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
Setting the output array as X2 i Output array Y2 i Output array Z2 i The values in the same sequence are added to obtain an input array Q of an output layer Z3 of the main task Z i ；
The Sigmoid function is also called a Logistic function, is also called an S-shaped growth curve, is used for hidden layer neuron output, has a value range of (0, 1), can map a real number to a section of (0, 1), can be used for classification, and has the advantages of smoothness and easiness in derivation when the Sigmoid is used as an activation function. The invention only adopts Sigmoid as the activation function of the neural network, and the principle of the Sigmoid function is not described here.
According to the invention, the output of the hidden layer of the auxiliary task is fused with the output of the hidden layer of the main task, and the fusion result is used as the input of the output layer of the main task, so that the weight sharing of the main task and the auxiliary task is realized, and the accuracy of whether IDH occurs at the next moment of main task prediction is effectively improved.
S4, training a model;
s41, constructing training set data and test set data;
sequencing the input data collected in the step S1 according to dialysis time, taking the first 80% as training set data and the last 20% as test set data;
s42, constructing a loss function;
firstly, constructing a loss function by adopting cross entropy for an auxiliary task X, an auxiliary task Y and a main task Z;
Then constructing a model total loss function, wherein the model total loss function is a weighted sum of the loss functions of the auxiliary task X, the auxiliary task Y and the main task Z; wherein,
the weight ratio of the cross entropy loss function of the main task Z to the auxiliary tasks X and Y is 2:1:1.
s43, calculating weight parameters of corresponding neurons when the loss function takes the minimum value by adopting a gradient descent method for training set data;
the neuron is a neuron in the model constructed in S3.
S5, verifying a multitasking model;
the multi-task model calculates the test set data obtained in the step S4 to obtain the results of three tasks of the test set data, calculates the accuracy rate, recall rate and precision rate of each task respectively, and evaluates the effect of the multi-task model through the three indexes.
The accuracy of the auxiliary task A is the ratio of the number of the model correctly predicted IDH occurrence and the IDH non-occurrence in the test set data and the total number of the test data;
the recall rate of the auxiliary task A is the ratio of the number of IDH occurrence correctly predicted by the model in the test set data to the total number of IDH occurrence in the test set data;
the accuracy of the auxiliary task A is the ratio of the number of IDH occurrence correctly predicted by the model in the test set data to the total number of IDH occurrence predicted by the model.
The accuracy of the auxiliary task B is the ratio of the sum of the numbers of the model correct prediction without intervention, suspension ultrafiltration, volume reduction, physiological saline infusion and up-regulation conductivity in the test set data to the total number of the test data;
the recall rate of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspension ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the non-intervention, the suspension ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data respectively;
the accuracy of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the model predicted non-intervention, the total number of the suspended ultrafiltration, the total number of the volume reduction, the total number of the physiological saline infusion and the total number of the up-regulation conductivity.
The accuracy of the main task C is the ratio of the number of the occurrence of IDH at the next moment and the non-occurrence of IDH at the next moment and the total number of the test data, which are correctly predicted by the model in the test set data;
The recall rate of the main task C is the ratio of the number of IDH occurrence at the next moment to the total number of IDH occurrence at the next moment in the test set data, which is correctly predicted by the model in the test set data;
the accuracy of the main task C is the ratio of the number of IDH occurrence at the next moment predicted by the model correctly in the test set data to the total number of IDH occurrence at the next moment predicted by the model.
The invention discloses a method for recommending a multitask model based on multitask learning through midrange hypotension prediction and intervention measures, which reduces the risk of IDH occurrence of a patient in a mode of early prediction and early intervention, thereby achieving the purposes of preventing and treating IDH in the MHD treatment process and improving the prognosis of the MHD patient, and specifically:
in one aspect, it is not only possible to predict whether IDH is occurring in the present dialysis prior to patient hemodialysis, to increase the medical care's attention to patients who are likely to develop IDH.
On the other hand, the model can also be used for predicting whether IDH occurs at the next monitoring moment in the dialysis process according to the monitored sign and dialysis treatment parameters during hemodialysis. Thirdly, by adopting the model of the invention, a high-quality IDH intervention scheme can be provided for medical staff, and a high-quality treatment scheme recommendation, namely a high-quality IDH intervention measure recommendation, is provided when IDH occurs in hemodialysis, so that the problems that the medical staff selects an improper treatment scheme or can not give the treatment scheme in a short time or the medical staff without experience does not know how to treat the patient, such as dizziness, dysphoria, anxiety, pale complexion, yawning, nausea, vomiting, chest distress, heart rate increase, abdominal discomfort, cold sweat, dyspnea, blackness, muscle spasm and even transient consciousness loss are solved, and acute cardiovascular events, even death and the like, are caused when serious patients are serious.
Fourth, the invention fuses the output of the hidden layer of the auxiliary task and the output of the hidden layer of the main task, and the fusion result is used as the input of the output layer of the main task, thereby realizing the weight sharing of the main task and the auxiliary task.
Example 2: the embodiment provides an application of an IDH prediction and intervention recommended multitasking model constructed by adopting the method of the invention in predicting IDH in maintenance hemodialysis treatment.
Before a patient starts hemodialysis, personal information, medical history records, inspection and examination results, historical dialysis records and the current hemodialysis prescription of the patient are called from the system, and physical sign data of the patient are measured to obtain time-invariant data before hemodialysis: pre-dialysis weight, pre-dialysis heart rate, pre-dialysis dry weight, pre-dialysis pulse, pre-systolic pressure, pre-diastolic pressure, pre-dialysis respiratory rate, age, sex, height, age of dialysis, history of hypertension, history of diabetes, blood white count, hemoglobin, blood cell count, blood glucose, blood albumin, total blood cholesterol, blood triglycerides, hemoglobin, uric acid, blood urea nitrogen, blood potassium, blood sodium, blood calcium, blood phosphorus, blood chlorine, urine white blood cells, urine proteins, urine red blood cells, urine creatinine, urine occult blood, urine microalbumin, urine albumin, weight gain during the dialysis interval, weight of last dialysis, number of IDH occurrences in the last seven days, number of IDH occurrences in the last thirty days, dialysis mode, anticoagulant dose, ultrafiltration volume, dialysis duration, dialysate potassium ion concentration, dialysate calcium ion concentration, dialysate sodium ion concentration, dialysate conductivity, blood flow volume;
The time-invariant data before hemodialysis is input into an IDH prediction and intervention recommendation multitasking model, which predicts the risk of IDH occurring in the present dialysis before the patient starts hemodialysis.
In the maintenance hemodialysis treatment of a patient, reading interval monitoring signs, dialysis treatment parameters and dialysis machine parameters which are obtained by continuous fixed time interval monitoring in a hemodialysis machine, and obtaining time-varying data at each time point during the hemodialysis: current body temperature, current heart rate, current systolic pressure, current diastolic pressure, current pulse, current respiratory rate, current ultrafiltered, current dialysis duration, current dialysate potassium ion concentration, current dialysate calcium ion concentration, current dialysate sodium ion concentration, current dialysate conductivity, current blood flow, current arterial pressure, current venous pressure, current cross-molding, current blood filtration volume.
The time-invariant data before hemodialysis and the time-varying data at each moment during hemodialysis are input into an IDH prediction and intervention recommendation multitasking model, and the model can predict whether IDH occurs at the next moment and the intervention recommendation of IDH at the next moment during the current dialysis of a patient.
Claims (7)
1. The method for constructing the hypotension IDH prediction and intervention recommended multitasking model in dialysis is characterized by comprising the following steps:
S1, collecting input data; collecting time-invariant data before hemodialysis and time-varying data at each time during hemodialysis as input data; forming an input data by each time-invariant data and each time-variant data together;
s2, setting a label for each piece of input data; setting 3 labels for each piece of input data, namely a label A, a label B and a label C; each label corresponds to a learning task;
the label A is: whether hypotension IDH in dialysis occurs in the present dialysis;
the label B is: whether hypotension IDH in dialysis occurs at the next moment;
the label C is: intervention measures of hypotension IDH in dialysis at the next moment; among these, interventions for hypotensive IDH in dialysis include: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity;
s3, constructing a hypotension prediction and intervention recommendation multitasking model in dialysis; the multi-task model comprises an auxiliary task X, an auxiliary task Y and a main task Z; the model framework of each task is equally divided into an input layer, a hidden layer and an output layer, and each layer is composed of a plurality of neurons; each neuron in the hidden layer and the output layer has a weight, and the weight is obtained in the process of model training; each neuron of the output layer can have different weight calculations on the data output by the hidden layer, and the final result is output, wherein the number of the neurons of the output layer is determined by the number of the label categories;
S4, training a model; constructing training set data and test set data from the input data collected in the step S1, constructing a loss function, inputting the training set data into a model, and training the model;
s5, verifying a multitasking model; the multi-task model calculates the test set data obtained in the step S4 to obtain the results of three tasks of the test set data, calculates the accuracy rate, recall rate and precision rate of each task respectively, evaluates the effect of the multi-task model through the three indexes, and obtains the qualified in-dialysis hypotension IDH prediction and intervention recommended multi-task model when the accuracy rate, recall rate and precision rate of each task reach preset values;
step S3, predicting whether the hypotension IDH in the dialysis occurs in the present dialysis as an auxiliary task X; predicting an intervention measure of hypotension IDH in dialysis at the next moment as an auxiliary task Y; predicting whether hypotension IDH in dialysis occurs at the next moment as a main task Z;
the model framework of the auxiliary task X is divided into an input layer X1, a hidden layer X2 and an output layer X3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer X2 and the output layer X3 has a weight which is obtained in the model training process;
5.1 auxiliary task X model includes the following structure:
(1) Input layer: the data input by the input layer X1 of the auxiliary task X are time-invariant data before hemodialysis, and the number of neurons of the input layer X1 is time-invariant data before hemodialysis;
(2) Hidden layer: each neuron of the hidden layer X2 may have different weight calculations on the data input by the input layer X1, so as to further favor the prediction of a certain task tag; in the invention, the weight value of the hidden layer X2 of the auxiliary task X is biased to predict whether the hypotension IDH in dialysis occurs in the present dialysis;
setting:
the output array of the input layer X1 of the auxiliary task X is X1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer X2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer X2 be X2i, then:
；
(3) Output layer: one neuron in the output layer X3, 0 is that the low blood pressure IDH in dialysis does not occur in the present dialysis, 1 is that the low blood pressure IDH in dialysis occurs in the present dialysis; the neuron of the output layer X3 has different weight calculations on the data output by the hidden layer, and outputs a final result;
Setting:
the neuron of the output layer X3 comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then, output of output layer X3 = Sigmoid # -)；
5.2 model of auxiliary task Y includes the following structure:
the model framework of the auxiliary task Y is divided into an input layer Y1, a hidden layer Y2 and an output layer Y3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Y2 and the output layer Y3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer Y1 of the auxiliary task Y is the same as the data input by the input layer of the main task Z1, and is time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Y1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period;
(2) Hidden layer: each neuron of the hidden layer Y2 may have different weight calculation on the data input by the input layer Y1, so as to further favor the prediction of a certain task label;
setting:
the output array of the input layer Y1 of the auxiliary task Y is Y1 i I is the number of numerical values in the output array, and i is 1-n;
the neuron of the hidden layer Y2 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
let the output data array of the hidden layer Y2 be Y2 i Then:
；
in the present invention, the weight value of the hidden layer Y2 of the auxiliary task Y is biased towards predicting the intervention of hypotension IDH in dialysis at the next moment;
(3) Output layer: five output layer Y3 neurons respectively correspond to intervention measures of low blood pressure IDH in dialysis at the next moment without intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity and the like;
the neuron of the output layer Y3 has different weight calculations on the data output by the hidden layer and outputs a final result;
setting:
the neuron of the output layer Y3 comprises a plurality of weight parameter arrays W j The number of the weight parameters W is j, and j is 1-n;
weight parameter array W j Number and weight parameter array W i Equal in number, i.e. i=j
Then, the output of the output layer Y3=sigmoid #)；
5.3 the main task Z model comprises the following structure:
the model framework of the main task Z is divided into an input layer Z1, a hidden layer Z2 and an output layer Z3, each layer is composed of a plurality of neurons, and each neuron in the hidden layer Z2 and the output layer Z3 has a weight which is obtained in the model training process;
(1) Input layer: the data input by the input layer of the main task Z1 is the same as the data input by the input layer Y1 of the auxiliary task Y, and is time-invariant data before hemodialysis and time-variant data at each time in the hemodialysis period, and the number of neurons of the input layer Z1 is the sum of the number of time-invariant data before hemodialysis and the number of time-variant data at each time in the hemodialysis period;
(2) Hidden layer: each neuron of the hidden layer Z2 may have different weight calculation on the data input by the input layer Z1, so that prediction of a certain task label is more favored, in the present invention, the weight value of the hidden layer Z2 of the main task Z is favored to predict whether hypotension IDH in dialysis occurs at the next moment;
(3) Output layer: one output layer Z3 neuron of the main task Z is as follows: the next time the hypotension IDH in dialysis does not occur, indicated by "0", or the next time the hypotension IDH in dialysis occurs, indicated by "1";
taking an output array of a hidden layer X2 of an auxiliary task X, an output array of a hidden layer Y2 of an auxiliary task Y and an output array of a hidden layer Z2 of a main task Z as inputs of an output layer Z3 of the main task Z;
setting:
hidden layer X2 of auxiliary task X The output array is X2 i I is the number of numerical values in the output array, and i is 1-n;
output array Y2 of hidden layer Y2 of auxiliary task Y i I is the number of numerical values in the output array, and i is 1-n;
the output array of the hidden layer Z2 of the main task Z is Z2 i I is the number of numerical values in the output array, and i is 1-n;
the number of the output array of the hidden layer X2 of the auxiliary task X, the number of the output array of the hidden layer Y2 of the auxiliary task Y and the number of the numerical values in the output array of the hidden layer Z2 of the main task Z are the same, i is 1-n;
the neuron of the main task Z output layer Z3 comprises a plurality of weight parameter arrays W i The number of the weight parameters W is i, and i is 1-n;
the number of the weight parameters W is equal to the number i of the numerical values in the output array;
setting the output array as X2 i Output array Y2 i Output array Z2 i The values in the same sequence are added to obtain an input array Q of an output layer Z3 of the main task Z i ；
Output of master task Z = Sigmoid @)。
2. The method for constructing a model for prediction and intervention recommendation multitasking for hypotension IDH in dialysis according to claim 1, wherein:
the unchanged data in step S1 includes: measured sign data, personal information, medical history records, test results, historical dialysis records, current hemodialysis orders, prior to patient hemodialysis; wherein,
Measuring the sign data includes: anterior body weight, anterior heart rate, anterior dry body weight, anterior pulse, anterior systolic pressure, anterior diastolic pressure, and anterior respiratory rate;
the personal information includes: age, sex, height, dialysis age;
the medical history record includes: a history of hypertension and diabetes;
the checking result includes: blood white count, hemoglobin, red blood cell count, blood glucose, blood albumin, blood total cholesterol, blood triglycerides, blood creatinine, blood uric acid, blood urea nitrogen, blood potassium, blood sodium, blood calcium, blood phosphorus, blood chlorine, urine white blood cells, urine proteins, urine red blood cells, urine creatinine, urine occult blood, urine microalbumin, urine albumin;
the history dialysis record includes: weight gain during dialysis, last dialysis time, last time weight, number of times of occurrence of hypotensive IDH in dialysis in seven days, and number of times of occurrence of hypotensive IDH in dialysis in thirty days;
the current hemodialysis prescription includes: dialysis mode, anticoagulant dose, ultrafiltration volume, dialysis duration, dialysate potassium ion concentration, dialysate calcium ion concentration, dialysate sodium ion concentration, dialysate conductivity, blood flow;
the time-varying data includes: during hemodialysis of a patient, monitoring physical signs, dialysis treatment parameters and dialysis machine parameters at fixed time intervals; wherein,
The interval monitoring sign data includes: current body temperature, current heart rate, current systolic pressure, current diastolic pressure, current pulse, current respiratory rate;
dialysis treatment parameters included: current ultrafiltered, current dialysis duration, current dialysate potassium ion concentration, current dialysate calcium ion concentration, current dialysate sodium ion concentration, current dialysate conductivity, current blood flow;
the dialysis machine parameters included: current arterial pressure, current venous pressure, current cross-compression molding, current blood flow.
3. The method for constructing a model for prediction and intervention recommendation multitasking for hypotension IDH in dialysis according to claim 1, wherein: step S2 comprises the following sub-steps:
s21, setting a diagnosis standard of hypotension IDH in dialysis:
(1) the intervention measure of hypotension IDH in dialysis is met, and the systolic pressure is reduced by more than 20mmHg compared with the systolic pressure before dialysis;
(2) no intervention, but a systolic blood pressure of less than 90mmHg;
s22, setting a label A, a label B and a label C for each piece of input data according to the diagnosis standard of hypotension IDH in dialysis.
4. The method for constructing a model for prediction of hypotension IDH in dialysis and recommendation for intervention according to claim 3, wherein: step S22 comprises the following sub-steps:
S221, collecting the hypotensive IDH intervention data in dialysis, wherein the hypotensive IDH intervention data in dialysis is an intervention measure of the input data in the dialysis process; the intervention measures of the hypotension IDH intervention data in dialysis are also as follows: 5 categories of no intervention, suspension of ultrafiltration, volume reduction, physiological saline infusion, up-regulation of conductivity;
s222, setting a label A; judging whether the input data of all times of the dialysis is the hypotensive IDH in the dialysis according to the diagnosis standard of the hypotensive IDH in the dialysis;
s223, setting a label B; judging whether the input data of all times of the dialysis is the hypotensive IDH in the dialysis according to the diagnosis standard of the hypotensive IDH in the dialysis; the label B at the moment of judging that the hypotension IDH in dialysis is the hypotension IDH in dialysis at the next moment, and the labels B at the rest moments are marked at the next moment and do not generate the hypotension IDH in dialysis;
s224, setting a label C; the input data acquisition tag C is set according to the intervention data of hypotension IDH in dialysis.
5. The method for constructing a model for prediction and intervention recommendation multiplexing of hypotension IDH under dialysis according to claim 1, wherein step S4 comprises the sub-steps of:
S41, constructing training set data and test set data;
sequencing the input data collected in the step S1 according to dialysis time, taking the first 80% as training set data and the last 20% as test set data;
s42, constructing a loss function;
firstly, constructing a loss function by adopting cross entropy for an auxiliary task X, an auxiliary task Y and a main task Z;
then constructing a model total loss function, wherein the model total loss function is a weighted sum of the loss functions of the auxiliary task X, the auxiliary task Y and the main task Z; wherein,
the weight ratio of the cross entropy loss function of the main task Z to the auxiliary tasks X and Y is 2:1:1, a step of;
s43, calculating weight parameters of corresponding neurons when the loss function takes the minimum value by adopting a gradient descent method for training set data;
the neuron is a neuron in the model constructed in S3.
6. The method for constructing a model for prediction and intervention recommendation multiplexing of hypotension IDH under dialysis according to claim 1, wherein in step S5:
the accuracy of the auxiliary task A is the ratio of the number of the model correctly predicting the occurrence of the hypotension IDH in dialysis to the non-occurrence of the hypotension IDH in dialysis in the test set data and the total number of the test data;
The recall rate of the auxiliary task A is the ratio of the number of occurrence of hypotensive IDH in dialysis, which is predicted by the model correctly in the test set data, to the total number of occurrence of hypotensive IDH in dialysis in the test set data;
the accuracy of the auxiliary task A is the ratio of the number of the occurrence of the hypotension IDH in the dialysis predicted by the model to the total number of the occurrence of the hypotension IDH in the dialysis predicted by the model in the test set data;
the accuracy of the auxiliary task B is the ratio of the sum of the numbers of the model correct prediction without intervention, suspension ultrafiltration, volume reduction, physiological saline infusion and up-regulation conductivity in the test set data to the total number of the test data;
the recall rate of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data respectively;
the accuracy of the auxiliary task B is the ratio of the number of the model correctly predicted non-intervention, the suspended ultrafiltration, the volume reduction, the physiological saline infusion and the up-regulation conductivity in the test set data to the total number of the model predicted non-intervention, the total number of the suspended ultrafiltration, the total number of the volume reduction, the total number of the physiological saline infusion and the total number of the up-regulation conductivity;
The accuracy of the main task C is the ratio of the number of the hypotension IDH occurrence in the dialysis at the next moment to the number of the hypotension IDH non-occurrence in the dialysis at the next moment and the total number of the test data, which are correctly predicted by the model in the test set data;
the recall rate of the main task C is the ratio of the number of occurrence of hypotension IDH in dialysis at the next moment to the total number of occurrence of hypotension IDH in dialysis at the next moment in the test set data, which is correctly predicted by the model in the test set data;
the accuracy of the primary task C is the ratio of the number of occurrences of hypotension IDH in dialysis at the next time predicted by the model to the total number of occurrences of hypotension IDH in dialysis at the next time predicted by the model in the test set data.
7. The method for constructing a predicted and recommended multi-task model of in-dialysis hypotension IDH according to claim 1, characterized in that the predicted and recommended multi-task model of in-dialysis hypotension IDH constructed by the method of claim 1 is applied in the maintenance hemodialysis treatment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210711223.3A CN114937486B (en) | 2022-06-22 | 2022-06-22 | IDH prediction and intervention measure recommendation multi-task model construction method and application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210711223.3A CN114937486B (en) | 2022-06-22 | 2022-06-22 | IDH prediction and intervention measure recommendation multi-task model construction method and application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114937486A CN114937486A (en) | 2022-08-23 |
| CN114937486B true CN114937486B (en) | 2023-09-26 |
Family
ID=82868450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210711223.3A Active CN114937486B (en) | 2022-06-22 | 2022-06-22 | IDH prediction and intervention measure recommendation multi-task model construction method and application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114937486B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110507296A (en) * | 2019-08-12 | 2019-11-29 | 重庆大学 | A kind of acute hypotension mixing method for early warning based on LSTM network |
| CN111028913A (en) * | 2019-11-29 | 2020-04-17 | 北京工业大学 | Hemodialysis treatment scheme aid decision-making method |
| CN111460157A (en) * | 2020-04-01 | 2020-07-28 | 哈尔滨理工大学 | Cyclic convolution multitask learning method for multi-field text classification |
| CN111939353A (en) * | 2019-05-14 | 2020-11-17 | 吴元昊 | Construction method of prediction model of hypotensive event in hemodialysis |
| CN112107752A (en) * | 2019-06-20 | 2020-12-22 | 纬创资通股份有限公司 | Blood pressure prediction method and electronic device using same |
| CN113270192A (en) * | 2021-04-19 | 2021-08-17 | 浙江大学 | Hemodialysis risk prediction system based on combination of dynamic and static data and depth self-encoder |
| CN114091460A (en) * | 2021-11-24 | 2022-02-25 | 长沙理工大学 | Multitask Chinese entity naming identification method |
| CN114186711A (en) * | 2021-10-27 | 2022-03-15 | 中山大学 | Industrial raw material consumption prediction method based on multitask time sequence learning |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SG10201501889WA (en) * | 2010-03-15 | 2015-05-28 | Singapore Health Serv Pte Ltd | Method of predicting the survivability of a patient |
| DE102013108543A1 (en) * | 2013-08-07 | 2015-02-26 | B. Braun Avitum Ag | Apparatus and method for predicting intradialytic parameters |
| US20210193317A1 (en) * | 2019-12-20 | 2021-06-24 | Fresenius Medical Care Holdings, Inc. | Real-time intradialytic hypotension prediction |
-
2022
- 2022-06-22 CN CN202210711223.3A patent/CN114937486B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111939353A (en) * | 2019-05-14 | 2020-11-17 | 吴元昊 | Construction method of prediction model of hypotensive event in hemodialysis |
| CN112107752A (en) * | 2019-06-20 | 2020-12-22 | 纬创资通股份有限公司 | Blood pressure prediction method and electronic device using same |
| CN110507296A (en) * | 2019-08-12 | 2019-11-29 | 重庆大学 | A kind of acute hypotension mixing method for early warning based on LSTM network |
| CN111028913A (en) * | 2019-11-29 | 2020-04-17 | 北京工业大学 | Hemodialysis treatment scheme aid decision-making method |
| CN111460157A (en) * | 2020-04-01 | 2020-07-28 | 哈尔滨理工大学 | Cyclic convolution multitask learning method for multi-field text classification |
| CN113270192A (en) * | 2021-04-19 | 2021-08-17 | 浙江大学 | Hemodialysis risk prediction system based on combination of dynamic and static data and depth self-encoder |
| CN114186711A (en) * | 2021-10-27 | 2022-03-15 | 中山大学 | Industrial raw material consumption prediction method based on multitask time sequence learning |
| CN114091460A (en) * | 2021-11-24 | 2022-02-25 | 长沙理工大学 | Multitask Chinese entity naming identification method |
Non-Patent Citations (2)
| Title |
|---|
| 基于学习向量量化网络建立糖尿病并发症预测模型;李戈;靳立忠;;中国自然医学杂志(04);全文 * |
| 面向不完备数据的民航旅客流失预测模型;李国;袁闻;王怀超;;计算机工程与设计(10);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114937486A (en) | 2022-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4276834B2 (en) | Biological information and blood processing apparatus information management system | |
| Kusiak et al. | Predicting survival time for kidney dialysis patients: a data mining approach | |
| US8246567B2 (en) | Method of monitoring hypertensive haemodialysis patients | |
| Madihally | Principles of biomedical engineering | |
| US10437958B2 (en) | Device and method for monitoring the treatment of a patient | |
| EP1938848A1 (en) | Peritoneal membrane function test method, peritoneal membrane function test apparatus and peritoneal membrane function test program | |
| CN107729709A (en) | Adjusted in adaptive peritoneal dialysis session for overall session optimization | |
| US8123715B2 (en) | Planning method and apparatus for peritoneal dialysis and hemodialysis hybrid remedy | |
| CN106063701A (en) | Blood fluid removes the monitoring between treatment stage | |
| EP0027470A1 (en) | Computer-controlled medical care system. | |
| US20180043076A1 (en) | Intersession adaptive peritoneal dialysis fluid removal for multiple session optimization | |
| CN111297329B (en) | Method and system for predicting dynamic onset risk of cardiovascular complications of diabetics | |
| CN113270192A (en) | Hemodialysis risk prediction system based on combination of dynamic and static data and depth self-encoder | |
| CN117153336B (en) | Hemodialysis monitoring system and method based on hemodialysis machine | |
| CN114937486B (en) | IDH prediction and intervention measure recommendation multi-task model construction method and application | |
| CN108986920B (en) | Dialysis adequacy protein decomposition rate evaluation method and system | |
| Marano et al. | A new approach to bicarbonate addition during hemodialysis: Testing model predictions in a patient cohort | |
| CN112951419B (en) | Hemodialysis dry weight intelligent assessment device | |
| Leaning et al. | A system of models for fluid-electrolyte dynamics | |
| RU2693241C1 (en) | Method of treating renal insufficiency in domestic animals by applying haemodialysis procedure | |
| CN112216394A (en) | Blood glucose change trend prediction system based on self-adaptive optimization model | |
| CN115579108B (en) | AI hemodialysis scheme model modeling method and method for generating hemodialysis scheme | |
| US20230293787A1 (en) | Heart failure risk monitoring and hemoglobin level prediction for hemodialysis systems | |
| CN113057586B (en) | Disease early warning method, device, equipment and medium | |
| CN113241174A (en) | Checking system for CRRT treatment parameters of blood purification patient |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: The construction method and application of a multi task model for IDH prediction and intervention recommendation Granted publication date: 20230926 Pledgee: Nanjing Bank Co.,Ltd. Nanjing North Branch Pledgor: Shentai Health Technology (Nanjing) Co.,Ltd. Registration number: Y2024980005062 |