Closed CraigOpie closed 1 year ago
This document outlines the technical breakdown for designing a gRPC API schema to allow secure communication and access control between clients and the server for healthcare data exchange. The schema uses mTLS RSA-2048 or greater cryptography for data in transit and client verification using a whitelist. The API follows the HL7 FHIR API standard for healthcare and uses gRPC instead of RESTful APIs.
The gRPC API schema should include:
Additionally, the following requirements and clarifications are necessary:
The following Protobuf message types are defined in the gRPC API schema for healthcare data exchange:
syntax = "proto3";
package healthcare;
option java_multiple_files = true;
option java_package = "com.example.healthcare";
option java_outer_classname = "HealthcareProto";
message Identifier {
string use = 1;
string system = 2;
string value = 3;
}
message Code {
string system = 1;
string code = 2;
string display = 3;
}
message Coding {
Code code = 1;
}
message CodeableConcept {
repeated Coding coding = 1;
}
message Quantity {
double value = 1;
string unit = 2;
string system = 3;
string code = 4;
}
message Ratio {
Quantity numerator = 1;
Quantity denominator = 2;
}
message Period {
string start = 1;
string end = 2;
}
message Reference {
string reference = 1;
}
message HumanName {
string family = 1;
repeated string given = 2;
}
message Patient {
string id = 1;
repeated Identifier identifier = 2;
repeated HumanName name = 3;
string gender = 4;
string birthDate = 5;
}
message Practitioner {
string id = 1;
repeated Identifier identifier = 2;
repeated HumanName name = 3;
string gender = 4;
string birthDate = 5;
repeated Qualification qualification = 6;
}
message Qualification {
CodeableConcept code = 1;
Period period = 2;
}
message Encounter {
string id = 1;
string status = 2;
CodeableConcept class = 3;
Reference subject = 4;
Period period = 5;
}
message Observation {
string id = 1;
string status = 2;
CodeableConcept code = 3;
Reference subject = 4;
string effectiveDateTime = 5;
Quantity valueQuantity = 6;
}
message Condition {
string id = 1;
Reference subject = 2;
CodeableConcept code = 3;
string onsetDateTime = 4;
CodeableConcept clinicalStatus = 5;
CodeableConcept verificationStatus = 6;
}
message Medication {
string id = 1;
CodeableConcept code = 2;
string status = 3;
repeated Ingredient ingredient = 4;
CodeableConcept form = 5;
}
message Ingredient {
CodeableConcept itemCodeableConcept = 1;
Ratio strength = 2;
}
service HealthcareService {
// Patient
rpc CreatePatient(CreatePatientRequest) returns (CreatePatientResponse);
rpc GetPatient(GetPatientRequest) returns (GetPatientResponse);
rpc UpdatePatient(UpdatePatientRequest) returns (UpdatePatientResponse);
rpc DeletePatient(DeletePatientRequest) returns (DeletePatientResponse);
// Practitioner
rpc CreatePractitioner(CreatePractitionerRequest) returns (CreatePractitionerResponse);
rpc GetPractitioner(GetPractitionerRequest) returns (GetPractitionerResponse);
rpc UpdatePractitioner(UpdatePractitionerRequest) returns (UpdatePractitionerResponse);
rpc DeletePractitioner(DeletePractitionerRequest) returns (DeletePractitionerResponse);
// Encounter
rpc CreateEncounter(CreateEncounterRequest) returns (CreateEncounterResponse);
rpc GetEncounter(GetEncounterRequest) returns (GetEncounterResponse);
rpc UpdateEncounter(UpdateEncounterRequest) returns (UpdateEncounterResponse);
rpc DeleteEncounter(DeleteEncounterRequest) returns (DeleteEncounterResponse);
// Observation
rpc CreateObservation(CreateObservationRequest) returns (CreateObservationResponse);
rpc GetObservation(GetObservationRequest) returns (GetObservationResponse);
rpc UpdateObservation(UpdateObservationRequest) returns (UpdateObservationResponse);
rpc DeleteObservation(DeleteObservationRequest) returns (DeleteObservationResponse);
// Condition
rpc CreateCondition(CreateConditionRequest) returns (CreateConditionResponse);
rpc GetCondition(GetConditionRequest) returns (GetConditionResponse);
rpc UpdateCondition(UpdateConditionRequest) returns (UpdateConditionResponse);
rpc DeleteCondition(DeleteConditionRequest) returns (DeleteConditionResponse);
// Medication
rpc CreateMedication(CreateMedicationRequest) returns (CreateMedicationResponse);
rpc GetMedication(GetMedicationRequest) returns (GetMedicationResponse);
rpc UpdateMedication(UpdateMedicationRequest) returns (UpdateMedicationResponse);
rpc DeleteMedication(DeleteMedicationRequest) returns (DeleteMedicationResponse);
}
The API endpoints for creating, retrieving, updating, and deleting patient records and medical information are defined as follows:
service HealthcareService {
rpc CreatePatient(Patient) returns (Patient) {}
rpc GetPatientById(string) returns (Patient) {}
rpc UpdatePatient(Patient) returns (Patient) {}
rpc DeletePatient(string) returns (google.protobuf.Empty) {}
rpc CreateMedicalInformation(MedicalInformation) returns (MedicalInformation) {}
rpc GetMedicalInformationById(string) returns (MedicalInformation) {}
rpc UpdateMedicalInformation(MedicalInformation) returns (MedicalInformation) {}
rpc DeleteMedicalInformation(string) returns (google.protobuf.Empty) {}
}
Access control is based on permissions defined in the Permission
message type. The permissions
field is included in the Patient
message type and is used to store the permissions for each practitioner in the system. The has_permission
helper function is used to check if a practitioner has the required permission for a given operation. The server-side implementation verifies the client's permissions before allowing them to perform any operation on a specific resource type.
def has_permission(practitioner, resource_type, operation):
for permission in practitioner.permissions:
if permission.resourceType == resource_type and permission.operation == operation:
return True
return False
In each RPC method, the has_permission
helper function is used to check the client's permissions before processing the request:
def CreatePatient(self, request, context):
practitioner = get_practitioner_from_context(context)
if not has_permission(practitioner, "Patient", "create"):
context.abort(grpc.StatusCode.PERMISSION_DENIED, "Permission denied")
# Proceed with creating the patient
# ...
Here is an example of an extended Practitioner
message type to include permissions:
message Practitioner {
string id = 1;
repeated Identifier identifier = 2;
repeated HumanName name = 3;
string gender = 4;
string birthDate = 5;
repeated Qualification qualification = 6;
repeated Permission permissions = 7;
}
The gRPC framework built-in error handling mechanisms are used to handle errors or unexpected requests.
The response format for the API requests is Protobuf. For each FHIR resource, there are create
, get
, update
, and delete
methods - each method requires a corresponding request and response message type.
// Patient
message CreatePatientRequest {
Patient patient = 1;
}
message CreatePatientResponse {
Patient patient = 1;
}
message GetPatientRequest {
string id = 1;
}
message GetPatientResponse {
Patient patient = 1;
}
message UpdatePatientRequest {
string id = 1;
Patient patient = 2;
}
message UpdatePatientResponse {
Patient patient = 1;
}
message DeletePatientRequest {
string id = 1;
}
message DeletePatientResponse {}
// Practitioner
message CreatePractitionerRequest {
Practitioner practitioner = 1;
}
message CreatePractitionerResponse {
Practitioner practitioner = 1;
}
message GetPractitionerRequest {
string id = 1;
}
message GetPractitionerResponse {
Practitioner practitioner = 1;
}
message UpdatePractitionerRequest {
string id = 1;
Practitioner practitioner = 2;
}
message UpdatePractitionerResponse {
Practitioner practitioner = 1;
}
message DeletePractitionerRequest {
string id = 1;
}
message DeletePractitionerResponse {}
// Encounter
message CreateEncounterRequest {
Encounter encounter = 1;
}
message CreateEncounterResponse {
Encounter encounter = 1;
}
message GetEncounterRequest {
string id = 1;
}
message GetEncounterResponse {
Encounter encounter = 1;
}
message UpdateEncounterRequest {
string id = 1;
Encounter encounter = 2;
}
message UpdateEncounterResponse {
Encounter encounter = 1;
}
message DeleteEncounterRequest {
string id = 1;
}
message DeleteEncounterResponse {}
// Observation
message CreateObservationRequest {
Observation observation = 1;
}
message CreateObservationResponse {
Observation observation = 1;
}
message GetObservationRequest {
string id = 1;
}
message GetObservationResponse {
Observation observation = 1;
}
message UpdateObservationRequest {
string id = 1;
Observation observation = 2;
}
message UpdateObservationResponse {
Observation observation = 1;
}
message DeleteObservationRequest {
string id = 1;
}
message DeleteObservationResponse {}
// Condition
message CreateConditionRequest {
Condition condition = 1;
}
message CreateConditionResponse {
Condition condition = 1;
}
message GetConditionRequest {
string id = 1;
}
message GetConditionResponse {
Condition condition = 1;
}
message UpdateConditionRequest {
string id = 1;
Condition condition = 2;
}
message UpdateConditionResponse {
Condition condition = 1;
}
message DeleteConditionRequest {
string id = 1;
}
message DeleteConditionResponse {}
// Medication
message CreateMedicationRequest {
Medication medication = 1;
}
message CreateMedicationResponse {
Medication medication = 1;
}
message GetMedicationRequest {
string id = 1;
}
message GetMedicationResponse {
Medication medication = 1;
}
message UpdateMedicationRequest {
string id = 1;
Medication medication = 2;
}
message UpdateMedicationResponse {
Medication medication = 1;
}
message DeleteMedicationRequest {
string id = 1;
}
message DeleteMedicationResponse {}
The server is responsible for verifying the client certificate. The server is configured to use mTLS for secure communication and client verification. The appropriate SSL/TLS credentials, including the server's private key, certificate, and trusted CA certificates, are set up in the server code. The server also verifies that the client's certificate is signed by a trusted CA and matches an entry in the server's whitelist.
The designed gRPC API schema for healthcare data exchange allows secure communication and access control between clients and the server. The schema uses mTLS RSA-2048 or greater cryptography for data in transit and client verification using a whitelist. The API follows the HL7 FHIR API standard for healthcare and uses gRPC instead of RESTful APIs. The schema includes API endpoints for creating, retrieving, updating, and deleting patient records and medical information. Data structures for patient records, medical information, and access control are defined. Access control is based on permissions, and the gRPC framework built-in error handling mechanisms are used to handle errors or unexpected requests. The response format for the API requests is Protobuf. The server is responsible for verifying the client certificate, and the appropriate SSL/TLS credentials are set up in the server code.
Description: Design the gRPC API schema for secure communication between clients and the server.