A cloud-oriented architecture for the remote assessment and follow-up of hospitalized patients

Abstract

During the last months the dramatic COVID-19 outbreak has exposed the fragility of our healthcare system, as well as the need for a smart remote follow-up system for the patients, in order to less the burden on the healthcare service and reduce the average hospitalization time. In this paper we proposed a solution designed to grant maximum flexibility by means of the allocation of resources on a cloud service for the remote follow-up of patients. Such resources can be used as a remote support for the caregiver both when planning or enforcing a therapeutic path. A major explanation behind follow-up regards the location and treatment of potentially adverse reactions after treatments. Physical side efects of the diferent modalities of treatment might be various and crippling after chemotherapy and radiotherapy. Moreover remote follow up can be a life-changing solution also on the economical side, due to the implication of therapeutic attendances for patients as far as missed work and travel costs that must likewise be comprehended in the overall economical burden. In an investigation of patients with testicular disease, Campbell et al. Finally such a solution could efectively improve the patient's adherence to the therapeutic plan. The ability to remotely follow follow-up is therefore a monetarily alluring choice as far as investment funds, also given the improved eficiencies, reduced cost and number of missed working days for the patient. Patients with a patient-held record may also take advantage of a more conscious and motivated interest over their own wellbeing, illness and treatment, with a direct impact on patient's adherence to the therapeutic plan.

Keywords

• eol>Cloud computing
• Quality of service
• Clinical management
• Follow up
• Psychological well-being

-

1. Introduction During the last months the dramatic COVID-19 outbreak has exposed the fragility of our healthcare system. Moreover it gave use the possibility to rethink diferent therapeutic procedures as well as cure protocols. The activities of the healthcare operators in hospital wards have been extensively remodeled, so as to ensure greater safety for the entire staf operating in the facility as well as to enforce the required sterilization protocols to ensure the patients well being. The hospital facilities had been required to enforce massive use of personal protective equipment, install sanitizing gel dispensers in every hallway and waiting room, determine a maximum limit of people at the same time in the same room, and, above all to apply social distancing. It follows that hospitalized patients have been subjected to limitations, in order to maintain the correct social distancing, involving their permanence as well of their relatives, also avoiding gatherings, and, sometime, limiting the entrances in the ward. The effect of these necessary limitations is to increase the isolation efect on the hospitalized patient. Therefore, while the patient follows a cure protocol, he must also be helped, with the same accuracy, by means of a parallel protocol that takes care also of the solitude experienced by the person.

Isolated patients are visited fewer times than nonisolated patients, moreover such isolated patients generally benefit of a shorter time span with their physicians. Because of the significantly lower contact time observed, particularly among the most severely ill of lfoor patients, a reexamination of the risk-benefit ratio of this infection control method has been proposed. In facts the attending physicians are about half as likely to examine patients in contact isolation compared with patients not in contact isolation [ 1 ]. Similarly, other studies have pointed out the concern that isolation may negatively afect not only the perceived quality of service but also the patients’ mental health [ 2, 3 ], with a substantial increase in anxiety and stress-related disorders [ 4, 5 ]. Finally [ 6 ] shows that isolation precautions are associated with adverse efects which may result in poorer hospital outcomes, a longer hospitalSU ONLINE SERVICE

CLOUD FRONTEND

BACKEND FRONTEND REMOTE CLIENT PATIENT’S

DATA

FRONTEND REMOTE

CLIENT CAREGIVER (USER)

TLN ONLINE

SERVICE EDITING

BACKEND HEALTHCARE

SERVICE (ADMIN)

TECHNICAL BACKEND TECHNICIAN (ADMIN) ization, an higher cost of care, as well as an higher rate update and management. The solution has been deof readmission to hospital within a month. signed to grant maximum flexibility allocating resour

Diferently from direct intervention, follow-up medi- ces on a cloud service. Such resources can be used as a cine does not base its protocol only on drugs and pre- remote support for the caregiver both when planning scriptions, on the contrary it build the intervention or enforcing a therapeutic path. around the patient [ 7, 8, 9, 10 ]. Trough standardization The paper is organized as follows. After this brief the caregivers are guided in making decisions regard- introduction, in the following Section 2 the designed ing the more appropriate therapeutic plan for a spe- system is described in its constituent parts. In Seccific conditions, while the medical practices can be ra- tion 3 we will focus on the management of the cloud tionalized improving, in the end, the general outcome services giving further details on the resource allocafor the therapy at full advantage of the patient’s well tion policies. Finally in Section 4 we will draw our being. Other fields of medicine can rely on very ef- conclusions. fective clinical prediction rules in order to reduce the uncertainty inherent the medical practice by defining how to use clinical findings to make predictions [ 11 ]. 2. The developed system Finally, it must be said that in certain cases it is uttermost dificult to draw methodology-proof clinical In Figure 1 a gross schema of the designed system is practice guidelines due to the extreme statistical and reported, this is composed by the following agents and subjective variability of the matter at hand [ 12 ]. components:

In the work presented on this paper we developed a label=. Frontend: unified cloud-based resource for the management and • Online interface execution of all the task related to the patient’s followup from the creation of the medical record, to its use, • Frontend remote clients

lbbel=. Backend: QOS HANDLER CLOUD VM CLOUD SERVICES

LSV

SSL

Local storage

HTTP caching Apache Local Cluster cache

Job Queque

DISTRIBUTED DB

LOG HANDLER

Local storage Slave DB Main DB USER DNS

NAMESERVER

2.1. Frontend • Administration backend

The frontend of the system has been developed by mea• Technical backend ns of the Angular JS [ 13, 14 ] framework in order to lcbel=. Cloud grant portability and compatibility with almost all the available hardware and software systems. In this man• Trusted Login Node (TLN) ner there are no particular requirements to interface • Cloud administration service (CAS) with the developed system, granted the ability to exe• Login nodes (LN) cute JavaScript on a browser-like application. Although a web browser would have suficed to interface with • Computing nodes (CN) the online service, we developed a simple ad-hoc ap• Storage Units (SU) plication to oversimplify the interface. In this manner it is possible to avoid unnecessary complication

The components are better described in the follow- for the caregivers that will benefit of this application. ing. The frontend remote client only provides the interface AMAZON LEX

IOT RULE

AMAZON API GATEWAY

AMAZON S3

AWS IOT CORE

AMAZON KINESIS

DATA STREAM

AMAZON PINPOINT

AMAZON DYNAMO DB

AMAZON DYNAMO DB

STREAMS for the final users, but it is not designed to create a new data, the the backend of the developed system provides task or operate for its standardization, since these lat- the necessary support for the design of new task and ter procedures are related with the backend. follow-up protocols as well as their standardization, ifnally it implements the tools for the technical ad2.2. Backend ministration of the overall platform. Diferently from the frontend, the backend provides two separated conWhile the frontend for the proposed system is con- soles for caregivers and technicians. The first allows stituted by the final interface that the users and pa- the caregiver to design a new task or protocol, insert tients can use for the only purpose of entering certain the item, provide the validation rules, and request to the system to analyze the validation data. The sec- Figure 3). ond console is reserved for technical administration in The resource request is provided to the cloud manterms of resource allocation, cloud management poli- ager component which uses the Amazon AWS APIs to cies, etc... efectively request the allocation of new virtual machines. The cloud manager interfaces with the AWS 2.3. Cloud IoT Core taking into consideration the AWS IoT rule component that determine the policies for the Amazon The cloud resources are allocated both for computa- Kinesis Data Stream. The Amazon Kinesis Data Stream tional and provisional purposes. Complete standard- is a real-time streaming service that provides eventized procedures can be used and implemented by means driven messaging and supports extended microservice of the frontend interface. In this case a simple set of architectures. This latter allows the processing requests queries can do the job, by selecting and extracting the trough the Amazon API Gateway once an admin has required data from the databases, distributed on sev- been logged and identified trough his credentials by eral storage units (SU), as well as by uploading data for the Amazon Lex component to access the Amazon S3 further use. On the other hand the standardization of service. a new test, due to the required statistical analysis, also In our system design also the database is distributed makes use of the computing nodes (CN). The cloud sys- on the cloud and supported by the Amazon DynamoDB tem is also provided with several login node in order services that allows data flow by means of the Amato avoid direct interactions with the allocated comput- zon DynamoDb Streams component. Data transactions ing nodes and storage units, excepted for the storage and session state are encrypted at-rest and securely units containing the public database useful to run the managed in the high-performance and scalable NoSQL fronted interface. Finally the technical administration datastore ofered by DynamoDB. of the cloud, due to the criticality of the matter, makes The Amazon DynamoDB Streams is also able to triguse of a trusted login node that ensure a grater security ger an AWS Lambda function in order to send notificalevel. The details on the cloud policies are given in the tions, by means of the Amazon Pinpoint and Amazon following Section 3. Polly services, to psychologist users when a patient has completed a test, as well as to send notification to a psychologist admin when the validation and stan3. The cloud environment dardization process advances or changes status. In our proposal the cloud environment is administered on a technical level by expert users. Moreover the sys- 4. Discussion and conclusion tem has been studied in order to adapt to the total load by allocating or freeing resources. For the im- In order to check the developed system on the field, plementation we supported our system with the Ama- we simulated the small scale implementation on an zon Web Services (AWS) [ 15 ], and particularly on the oncology ward of an hospital subjected to personnel AWS ECS and S3 service [ 16 ]. The main component and visitors restriction due to an epidemic outbreak. for the administration of the cloud environment is the In this scenario all members of the oncological staf Cloud Administration Service (CAS) depicted in Fig- have to carry out certain activities, where possible, ure 4. The Cloud Administration Service analyzes the in a smart working mode (e.g. the drafting of scienprovided application and estimate the computational tific articles, end-of-treatment reports and other inburden by means of a XML application requirement ternal reports, statistical analysis and supplementary descriptor. Along with the single application (e.g. a clinical reviews). While all the measures adopted in meta-analysis by means of factorial analysis, multidi- such a scenario are necessary, as public health is almensional scaling, etc...), the psychologist administra- ways at risk, these measures evidently dilute the eftor also submits a set of requirements (e.g. the desired fectiveness of the treatment also remanding to the padeadline or throughput, etc...). Both the application tient’s responsibilities to follow the planned path. In descriptor and the submitted requirements, are then such a scenario the remote follow-up of patients beanalyzed by the request handler module. The request came a mandatory measure to both enforce the neceshandler has the responsibility to determine the correct sary sanitary safety protocols, as well as maintain an allocation request mediating between the application high level of therapeutic quality in order to ensure the requirements, the user defined constraints, and the ef- patients’ adherence to their therapeutic paths. In the fective resource availability on the cloud system (see case study at hands, therefore, we propose to implement and use the developed system in order to allow a remote follow-up of the patients. In this manner the hospitalization time can be reduced while preserving the patient’s remission time and the therapeutic eficacy. A major explanation behind follow-up regards the location and treatment of potentially adverse reactions after treatments. In facts such reactions might be long lasting. Sometimes the need to inform the patient can be dificult to reconcile with the doctors’ desire to encourage the latter: constant collaboration with psychologists will therefore allow them to better address these issues. Psychology responds to the need for specific reflection on the psychic processes involved in the adaptation of patients to the disease and on the evaluation of their quality of life. It must therefore provide useful tools for organizing the training of all the professional figures involved and propose efective strategies in psychological support to the patient. It is a set of knowledge and a series of skills in constant evolution, on which the professional identity of the psychologist is based. Physical side efects of the diferent modalities of treatment might be various and crippling after chemotherapy and radiotherapy. Moreover remote follow up can be a life-changing solution also on the economical side, due to the implication of therapeutic attendances for patients as far as missed work and travel costs that must likewise be comprehended in the overall economical burden.

In an investigation of patients with testicular disease it was demonstrated that every center participation costs a normal of £61 per understanding for loss of profit, with an extra £11 in movement costs. A similar report indicated elevated levels of fulfillment, with a proposition to supplant up close and personal meetings with calls. Finally such a solution could efectively improve the patient’s adherence to the therapeutic plan. The ability to remotely follow follow-up is therefore a monetarily alluring choice as far as investment funds, also given the improved eficiencies, reduced cost and number of missed working days for the patient. Patients with a patient-held record may also take advantage of a more conscious and motivated interest over their own wellbeing, illness and treatment, with a direct impact on patient’s adherence to the therapeutic plan.

References

[1] S. Saint , L. A. Higgins , B. K. Nallamothu , C. Chenoweth , Do physicians examine patients in contact isolation less frequently? a brief report , American journal of infection control 31 ( 2003 ) 354 - 356 .

[2] K. B. Kirkland , J. M. Weinstein , Adverse efects of contact isolation , The Lancet 354 ( 1999 ) 1177 - 1178 .

[3] B. Guilley-Lerondeau , C. Bourigault , A. -C. G. des Buttes , G. Birgand, D. Lepelletier , Adverse effects of isolation: a prospective matched cohort study including 90 direct interviews of hospitalized patients in a french university hospital , European Journal of Clinical Microbiology & Infectious Diseases 36 ( 2017 ) 75 - 80 .

[4] G. Capizzi , C. Napoli , S. Russo , M. Wozniak , Lessening stress and anxiety-related behaviors by means of ai-driven drones for aromatherapy , in: Proceedings of the 6th Italian Workshop on Artificial Intelligence and Robotics co-located with the XVIII International Conference of the Italian Association for Artificial Intelligence (AIxIA 2019 ), volume 2594 , CEUR-WS , 2019 , pp. 7 - 12 .

[5] G. Catalano , S. H. Houston , M. C. Catalano , A. S. Butera , S. M. Jennings , S. M. Hakala , S. L. Burrows , M. G. Hickey , C. V. Duss , D. N. Skelton , et al., Anxiety and depression in hospitalized patients in resistant organism isolation ., Southern medical journal 96 ( 2003 ) 141 - 146 .

[6] K. Tran , C. Bell , N. Stall , G. Tomlinson , A. McGeer , A. Morris , M. Gardam , H. B. Abrams , The efect of hospital isolation precautions on patient outcomes and cost of care: a multi-site, retrospective, propensity score-matched cohort study , Journal of general internal medicine 32 ( 2017 ) 262 - 268 .

[7] S. Russo , C. Napoli , A comprehensive solution for psychological treatment and therapeutic path planning based on knowledge base and expertise sharing , in: "Proceedings of ICYRIME 2019" , volume 2472 , CEUR-WS , 2019 , pp. 41 - 47 .

[8] M. Wózniak , D. Połap , R. K. Nowicki , C. Napoli , G. Pappalardo , E. Tramontana, Novel approach toward medical signals classifier , in: 2015 International Joint Conference on Neural Networks (IJCNN) , IEEE, 2015 , pp. 1 - 7 .

[9] C. Napoli , G. Pappalardo , E. Tramontana, R. K. Nowicki , J. T. Starczewski , M. Woźniak , Toward work groups classification based on probabilistic neural network approach , in: International Conference on Artificial Intelligence and Soft Computing , Springer, 2015 , pp. 79 - 89 .

[10] D. Połap , M. Woźniak , C. Napoli , E. Tramontana, Real-time cloud-based game management system via cuckoo search algorithm , International Journal of Electronics and Telecommunications 61 ( 2015 ) 333 - 338 .

[11] B. M. Reilly , A. T. Evans , Translating clinical research into clinical practice: impact of using prediction rules to make decisions , Annals of internal medicine 144 ( 2006 ) 201 - 209 .

[12] T. M. Shaneyfelt , M. F. Mayo-Smith , J. Rothwangl , Are guidelines following guidelines?: The methodological quality of clinical practice guidelines in the peer-reviewed medical literature , Jama 281 ( 1999 ) 1900 - 1905 .

[13] B. Green , S. Seshadri, AngularJS, " O'Reilly Media , Inc." , 2013 .

[14] P. B. Darwin , P. Kozlowski, AngularJS web application development, Packt Publ., 2013 .

[15] J. Murty , Programming amazon web services: S3, EC2, SQS, FPS, and SimpleDB , " O'Reilly Media , Inc." , 2008 .

[16] Z. Xie , Y. Chen , T. Jiang , J. Speer , T. Walters , P. A. Tarazaga , M. Kasarda , On-demand big data analysis in digital repositories: A lightweight approach , in: International Conference on Asian Digital Libraries, Springer, 2015 , pp. 274 - 277 .