Tatiana Ermakova

Collaborative and secure sharing of healthcare data in multi-clouds

In healthcare, inter-organizational sharing and collaborative use of big data become increasingly important. The cloud-computing paradigm is expected to provide an environment perfectly matching the needs of collaborating healthcare workers. However, there are still many security and privacy challenges impeding the wide adoption of cloud computing in this domain. In this paper, we present a novel architecture and its implementation for inter-organizational data sharing, which provides a high level of security and privacy for patient data in semi-trusted cloud computing environments. This architecture features attribute-based encryption for selective access authorization and cryptographic secret sharing in order to disperse data across multiple clouds, reducing the adversarial capabilities of curious cloud providers. An implementation and evaluation by several experiments demonstrate the practical feasibility and good performance of our approach.

Secret Sharing for Health Data in Multi-provider Clouds

The accelerated adoption of cloud computing among enterprises is due to the multiple benefits the technology provides, one of them the simplification of inter-organizational information sharing, which is of utmost importance in healthcare. Nevertheless, moving sensitive health records to the cloud still implies severe security and privacy risks. With this background, we present a novel secure architecture for sharing electronic health records in a cloud environment. We first conducted a systematic literature review and interviews with different experts from the German healthcare industry that allowed us to derive real-world processes and corresponding security and privacy requirements. Based on these results, we designed our multi-provider cloud architecture that satisfies many of the requirements by providing increased availability, confidentiality and integrity of the medical records stored in the cloud. This architecture features secret sharing as an important measure to distribute health records as fragments to different cloud services, which can provide higher redundancy and additional security and privacy protection in the case of key compromise, broken encryption algorithms or their insecure implementation. Finally, we evaluate and select a secret-sharing algorithm for our multi-cloud architecture. We implemented both Shamirs secret-sharing scheme and Rabins information dispersal algorithm and performed several experiments measuring the execution time. Our results indicate that an adoption of Rabins algorithm would create a low overhead, giving strong indicators to the feasibility of our approach.

Enabling location-based policies in a healthcare cloud computing environment

In a multi-stakeholder cloud computing environment, data access control is of essential importance. Nowadays, it is usually handled in and deployed by every single cloud service on its own which makes the configuration of fine-grained access privileges cumbersome and economically expensive. In this paper, we introduce a novel cloud ecosystem architecture featuring an overall lightweight data access control model. This model is enabling data access policies based on location information of service consumer devices. We apply our architecture in the sensitive healthcare domain, which itself comprises multiple parties with complex data access privileges. Here, we define high-level requirements driven from current data protection regulations and guidelines as well as practice requirements in this area, which we address in the design of our architecture. We implement and test the main components. The results demonstrate the feasibility of our architecture and the applicability of our approach even in the healthcare application domain.

Antecedents of Health Information Privacy Concerns

Abstract Due to demographic changes, health information technologies comprising electronic health records (EHR), electronic medical records (EMR), personal health records (PHR), remote patient monitoring (RPM) and other healthcare related websites are gaining significant relevance. They constitute a great opportunity for efficiency gains and further benefits. One of the major barriers to their successful adoption involves individual health information privacy concerns. In order to understand their nature and better mitigate them, this narrative literature survey deals with the antecedents of these concerns. In particular, this study identifies type of information, health status, recipient of information, knowledge of health information technology, experience of privacy invasions, age, gender, and education as highly important characteristics.

Phonomicrosurgery in Vocal Fold Nodules: Quantification of Outcomes in Professional and Non-Professional Voice Users

There are few data demonstrating the specific extent to which surgical intervention for vocal fold nodules (VFN) improves vocal function in professional (PVU) and non-professional voice users (NVU). The objective of this study was to compare and quantify results after phonomicrosurgery for VFN in these patient groups. METHODS In a prospective clinical study, surgery was performed via microlaryngoscopy in 37 female patients with chronic VFN manifestations (38±12 yrs, mean±SD). Pre- and postoperative evaluations of treatment efficacy comprised videolaryngostroboscopy, auditory-perceptual voice assessment, voice range profile (VRP), acoustic-aerodynamic analysis, and voice handicap index (VHI-9i). The dysphonia severity index (DSI) was compared with the vocal extent measure (VEM). RESULTS PVU (n=24) and NVU (n=13) showed comparable laryngeal findings and levels of suffering (VHI-9i 16±7 vs 17±8), but PVU had a better pretherapeutic vocal range (26.8±7.4 vs 17.7±5.1 semitones, p<0.001) and vocal capacity (VEM 106±18 vs 74±29, p<0.01). Three months postoperatively, all patients had straight vocal fold edges, complete glottal closure, and recovered mucosal wave propagation. The mean VHI-9i score decreased by 8±6 points. DSI increased from 4.0±2.4 to 5.5±2.4, and VEM from 95±27 to 108±23 (p<0.001). Both parameters correlated significantly (rs=0.82). The average vocal range increased by 4.1±5.3 semitones, and the mean speaking pitch lowered by 0.5±1.4 semitones. CONCLUSIONS These results confirm that phonomicrosurgery for VFN is a safe therapy for voice improvement in both PVU and NVU who do not respond to voice therapy alone. Top-level artistic capabilities in PVU were restored, but numeric changes of most vocal parameters were considerably larger in NVU.

Designing a Business Model for a Cloud Marketplace for Healthcare

Cloud platforms providing healthcare cloud services are supposed to start a new era in the healthcare industry. Nevertheless, appropriate marketplace business models have not been examined in this domain so far. This paper contributes by presenting a first business model draft for such a marketplace. Following the action research principles, we collaborate closely with hospital workers, software providers and developers. Based on the existing literature and findings from the TRESOR research project on the adoption of cloud computing in healthcare, by using the business model canvas we create a business model showing the main value proposition, the customer segments and the value architecture of the cloud marketplace. Furthermore the problem of multisided markets is addressed, as the business models show how it can be prevented.