Systems and methods for automated pharmaceutical stock management
The current disclosure relates to systems and methods for prescription stock management at one or more pharmacies in a pharmacy chain. The system initializes and trains a stock management model using machine learning techniques in order to calculate outputs such as a minimum stocked number, a maximum stocked number, and an order amount for at least one prescription item. The system then monitors the inventory of the prescription item, and transmits an order, based on the order amount, for the prescription item if the stock of the prescription item is less than the minimum stocked number.
1 . A computer-implemented method of performing improved prescription stock management, comprising:
initializing, using a plurality of machine learning algorithms, a stock management model for at least one prescription item of a plurality of prescription items,
wherein the stock management model produces forecasted outputs including a minimum stocked number, a maximum stocked number, an order amount for the at least one prescription item, and a threshold value for an inventory of the at least one prescription item,
wherein initializing the stock management model further includes calculating initial forecasted outputs each corresponding to a respective number of units of a respective prescription item,
wherein the initial forecasted outputs are based on forecasts for a predefined amount of time, and
wherein at least the order amount is variable based on the predefined amount of time;
determining, using the plurality of machine learning algorithms, one or more stock thresholds based on the forecasted outputs, wherein the one or more stock thresholds prevent understocking or overstocking prescription items;
monitoring, by one or more processors, an inventory for one or more prescription items from the plurality of prescription items at a pharmacy, wherein the monitoring includes:
(a) training, by one or more processors using a machine learning algorithm, the stock management model using training data,
wherein the machine learning algorithm includes one or more of: weighted regression, or blending, and
wherein the training includes improving accuracy of at least one output of the initial forecasted outputs by:
(1) receiving feedback data that indicates the accuracy of the initial forecasted outputs after a period of time including success of each order amount over the period of time; and
(2) adjusting, as additional training data and feedback is received, weights of the stock management model,
wherein the adjusting includes generating refined forecasted outputs corresponding to the initial forecasted outputs;
(b) determining that an inventory for at least one prescription item of the one or more prescription items satisfies a corresponding stock threshold; and
(c) transmitting, by one or more processors, an order to an external computing device when the inventory for the at least one prescription item is less than a corresponding minimum stocked number, wherein the order is based on the order amount; and
receiving, by one or more processors, one or more stock management algorithms;
receiving, by one or more processors, initialization data; and
calculating, by one or more processors and based on the one or more stock management algorithms and initialization data, initial values for the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
2 . The computer-implemented method of claim 1 , wherein training the stock management model further comprises:
receiving, by one or more processors, the training data corresponding to the at least one prescription item of the plurality of prescription items; and
recalculating, by one or more processors and based on the training data, the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
3 . The computer-implemented method of claim 1 , wherein the training data includes one or more of inventory data, customer data, pharmacy data, or historic data.
4 . The computer-implemented method of claim 1 , wherein the initial forecasted outputs and the forecasted outputs of the stock management model are based on seasonal demand.
5 . The computer-implemented method of claim 1 , wherein initialization data comprises data corresponding to a period of time, the period of time including one of at least a 104 week period, at least a 52 week period, or at least a 4 week period.
6 . The computer-implemented method of claim 1 , further comprising: causing, at a user-interface of a computer of the pharmacy, a display of an indication of a transmission of the order.
7 . A system for performing improved prescription stock management, the system comprising:
a database configured to store an inventory for one or more prescription items from a plurality of prescription items at a pharmacy,
an input device communicatively coupled to the database,
one or more non-transitory storage media configured to store processor-executable instructions, and
one or more processors operatively connected to the database, the input device, and the one or more non-transitory storage media and configured to execute the processor-executable instructions to cause the system to:
initialize, using a plurality of machine learning algorithms, a stock management model for at least one prescription item of a plurality of prescription items,
wherein the stock management model produces forecasted outputs including a minimum stocked number, a maximum stocked number, an order amount for the at least one prescription item, and a threshold value for an inventory of the at least one prescription item,
wherein to initialize the stock management model further includes to calculate initial forecasted outputs each corresponding to a respective number of units of a respective prescription item,
wherein the initial forecasted outputs are based on forecasts for a predefined amount of time, and
wherein at least the order amount is variable based on the predefined amount of time;
determine, using the plurality of machine learning algorithms, one or more stock thresholds based on the forecasted outputs, wherein the one or more stock thresholds prevent understocking or overstocking prescription items;
monitor an inventory for one or more prescription items from the plurality of prescription items at a pharmacy, wherein the monitoring includes:
(a) training, using a machine learning algorithm, the stock management model using training data,
wherein the machine learning algorithm includes one or more of: weighted regression, or blending, and
wherein the training includes improving accuracy of at least one output of the initial forecasted outputs by:
(1) receiving feedback data that indicates the accuracy of the initial forecasted outputs after a period of time including success of each order amount over the period of time; and
(2) adjusting, as additional training data and feedback is received, weights of the stock management model,
wherein the adjusting includes generating refined forecasted outputs corresponding to the initial forecasted outputs;
(b) determining that an inventory for at least one prescription item of the one or more prescription items satisfies a corresponding stock threshold; and
(c) transmitting an order to an external computing device when the inventory for the at least one prescription item from the plurality of prescription items is less than a corresponding minimum stocked number, wherein the order is based on the order amount;
receive one or more stock management algorithms;
receive initialization data; and
calculate, based on the one or more stock management algorithms and initialization data, initial values for the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
8 . The system of claim 7 , wherein initialization data comprises data corresponding to a period of time, the period of time including one of at least a 104 week period, at least a 52 week period, or at least a 4 week period.
9 . The system of claim 7 , wherein the instructions to train the stock management model further cause the system to:
receive the training data corresponding to the at least one prescription item of the plurality of prescription items; and
recalculate, based on the training data the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
10 . The system of claim 7 , wherein the at least one prescription item is one or more of a prescription medication, a non-prescription medication, or a medical device.
11 . The system of claim 7 , wherein the initial forecasted outputs and the forecasted outputs of the stock management model are based on seasonal demand.
12 . The system of claim 7 , wherein the input device further comprises a user-interface, and wherein the instructions further cause the system to:
display at the user-interface of the input device an indication of a transmission of the order.
13 . The system of claim 7 , wherein the system further includes:
one or more databases storing initialization data and training data, and wherein the one or more databases include inventory data, customer data, pharmacy data, and historic data.
14 . A tangible, non-transitory computer-readable medium storing processor-executable instructions that, when executed by one or more processors of a system, cause the system to:
initialize, using a plurality of machine learning algorithms, a stock management model for at least one prescription item of a plurality of prescription items,
wherein the stock management model produces forecasted outputs including a minimum stocked number, a maximum stocked number, an order amount for the at least one prescription item, a threshold value for an inventory of the at least one prescription item,
wherein to initialize the stock management model further includes to calculate initial forecasted outputs each corresponding to a respective number of units of a respective prescription item,
wherein the initial forecasted outputs are based on forecasts for a predefined amount of time, and
wherein at least the order amount is variable based on the predefined amount of time;
determine, using the plurality of machine learning algorithms, one or more stock thresholds based on the forecasted outputs, wherein the one or more stock thresholds prevent understocking or overstocking prescription items;
monitor, by one or more processors, an inventory for one or more prescription items from the plurality of prescription items at a pharmacy, wherein the monitoring includes:
(a) train, using a machine learning algorithm, the stock management model using training data,
wherein the machine learning algorithm includes one or more of: weighted regression, or blending, and
wherein the training includes improving accuracy of at least one output of the initial forecasted outputs by:
(1) receiving feedback data that indicates the accuracy of the initial forecasted outputs after a period of time including success of each order amount over the period of time; and
(2) adjusting, as additional training data and feedback is received, weights of the stock management model,
wherein the adjusting includes generating refined forecasted outputs corresponding to the initial forecasted outputs; and
(b) determining that an inventory for at least one prescription item of the one or more prescription items satisfies a corresponding stock threshold; and
(c) transmit an order to an external computing device when the inventory for the at least one prescription item is less than a corresponding minimum stocked number, wherein the order is based on the order amount;
receive one or more stock management algorithms;
receive initialization data; and
calculate, based on the one or more stock management algorithms and initialization data, initial values for the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
15 . The tangible, non-transitory computer-readable medium of claim 14 , wherein the processor-executable instructions, when executed, further cause the system to:
receive the training data corresponding to the at least one prescription item of the plurality of prescription items; and
recalculate, based on the training data the minimum stocked number, the maximum stocked number, the order amount for the at least one prescription item, and the threshold value.
16 . The tangible, non-transitory computer-readable medium of claim 14 , wherein the processor-executable instructions, when executed, further cause the system to:
display, at a user-interface of an input device, an indication of a transmission of the order.