The beginning of our collaboration

Victor Rențea was not one of my students. He came from the Faculty of Automatic Control and Computers, specialising in Computer Engineering, where he had acquired a solid education. I no longer remember exactly why he approached me in particular, although there were numerous doctoral supervisors in that faculty. His choice may also have been related to the fact that I was already firmly oriented toward a technological area that his faculty still represented rather weakly at the time: Java, J2EE, design patterns, software architectures, and, more generally, a way of thinking about software systems closer to their modern engineering construction.

Victor was a well-prepared and intelligent young man, eager to learn as much as possible about the field that interested him. At the time, my courses were in many respects at a higher level than anything he had previously encountered. I do not think he ever explicitly acknowledged this gap, but it is significant that he later devoted himself precisely to the technologies and topics that abounded in my courses: enterprise Java, architectures, patterns, and the clean design of systems. Looking back, the road from the young man he was then to the highly visible and self-assured trainer he is today has indeed been a long one.

I formulated the thesis topic, and he agreed to it. This marked the beginning of his true doctoral apprenticeship, because he had to study seriously everything I had already done in e-health: patient context, prevention, behaviour change, EHRs, HL7 standards, conceptual modelling, and the role of intelligent systems in supporting health. He possessed the necessary computing background, but the terrain of the thesis was broader and more complex than technology alone.

The doctoral thesis

Thesis: Contribuții la proiectarea unui sistem inteligent de suport al prevenției bolilor și al promovării sănătății
Contributions to the design of an intelligent support system for disease prevention and health promotion

Victor Rențea’s thesis begins from the observation that healthcare systems face a profound structural difficulty: the growth of chronic disease, population ageing, the increasing costs of curative medicine, and the limited effectiveness of an exclusively reactive approach. In this context, prevention and health promotion become not merely useful additions to traditional medicine, but a strategic direction for reorganising the relationship among the individual, the physician, and the healthcare system.

The structure of the thesis is stronger than may be apparent on a first reading. After the introductory chapter devoted to prevention, health promotion, and behaviour change, there follows a chapter on technologies relevant to e-health—HL7, EHRs, risk assessment, and multi-agent systems. The core of the work then moves to Chapters 3 and 4: modelling the prevention system, defining the Prevention Assistant System, and designing it as a multi-agent system. Finally, the thesis does not remain at the model level, but proceeds to a prototype and an experiment.

From my point of view, one of the thesis’s most fertile ideas is the treatment of prevention as a cybernetic feedback process. Victor starts from the traditional relationship between physician and patient and proposes adding a second loop around the individual, in which a software system monitors relevant data, estimates risks, and seeks to guide behaviour toward healthier patterns. In this way, prevention leaves the realm of general recommendations and enters a regime of continuous monitoring, response, and assistance.

The thesis’s second major direction is the modelling of the information required for prevention. Victor is not content with the traditional medical data found in an electronic record, but seeks to identify the determinants of health relevant to prevention and organise them within an ontology of his own. On this basis, he introduces the Virtual Individual Model, conceived as a complementary extension of the EHR and capable of including the person’s profile, risk factors, behaviours, preferences, physical activity, sleep, and other elements useful for preventive interventions.

The architectural core of the work is, however, the Prevention Assistant System, conceived as a multi-agent system. Here his training as a computer scientist and the influence of my own research directions are both clearly visible. The system is organised as a virtual prevention team around the individual, with distinct roles: monitoring parameters, assessing risks, constructing and following health-promotion initiatives, and supporting behaviour change. The notion of the Prevention Virtual Team is one of the thesis’s most interesting components because it connects conceptual modelling with a coherent software architecture.

Another merit of the thesis is its attempt to connect this architecture to real medical infrastructure rather than to an isolated application. For this reason, Victor explicitly discusses HL7 standards, the Clinical Document Architecture, and a prevention-oriented extension of the EHR. This choice gives engineering substance to the entire construction and brings it closer to the world of interoperable medical systems, which directly interested me as well.

Finally, the thesis also contains a significant experimental component. The implemented prototype and the experiment do not solve every problem or close the research definitively, but they show that the proposed ideas do not remain purely theoretical. Moreover, the experiment’s conclusions are interesting precisely because of their realism: the system appears useful, but behaviour change remains extremely difficult, confirming that prevention cannot be reduced to the simple delivery of messages, but requires a richer and more finely tuned interaction.

The thesis’s central ideas

Prevention as a cybernetic system

One of the thesis’s most interesting ideas is the treatment of prevention within a cybernetic feedback model. Victor starts from the idea of traditional control exercised by the physician over the patient’s condition and adds a second loop in which an intelligent system monitors the individual’s behaviour and intervenes to guide it toward healthier patterns. It is an elegant and fertile scheme because it moves prevention from the area of general advice into that of a monitored and assisted process.

Behaviour change and individual personality

The thesis assigns major importance to human behaviour, viewed as a decisive element in prevention. Victor discusses several behaviour-change models and ultimately focuses on the TTM (Trans-Theoretical Model), with its six stages. In addition, he attempts to correlate the process of change with personality type, using the MBTI. This shows his attempt to construct more personalised prevention and not to treat the individual merely as an abstract bearer of risks.

Conceptual modelling of prevention

A substantial contribution of the thesis lies in modelling. Victor attempts to define a taxonomy and ontology of prevention, identify the types of data relevant to prevention, and construct a virtual model of the individual. This virtual model brings together information on personal profile, risk factors, behaviours, preferences, physical activity, sleep, and mental state. From my point of view, this is one of the thesis’s best intuitions, even though this part might have deserved still greater development.

The Prevention Assistant System and the multi-agent paradigm

The architectural core of the thesis is the Prevention Assistant System, conceived as a multi-agent system. The idea is for different agents to play distinct roles in supporting prevention: assessing risks, constructing health-promotion plans, monitoring progress, and interacting with the individual. The system is conceived as a virtual prevention team built on the principle of virtualising actors and functions from the real world.

Here the profile of the computer scientist trained in Automatic Control and Computers and oriented toward architectures is particularly clear. Victor succeeds in connecting conceptual modelling to a well-articulated software architecture, using multi-agent systems and the Gaia methodology appropriately.

EHR, HL7, and prevention sections

Another interesting line of the thesis is the attempt to integrate prevention into the existing medical information infrastructure. Rather than treating prevention as an isolated application, Victor attempts to extend the electronic health record with sections specifically devoted to prevention and to use mechanisms compliant with HL7-CDA standards. This is an important direction because it gives the idea engineering consistency and removes it from the area of demonstrative applications disconnected from the real medical ecosystem.

Publications associated with the thesis period

Four peer-reviewed conference papers formed around Victor’s thesis, three of which, as I also noted in my report, were ISI-indexed. These works reflect the principal directions of his doctoral research quite well: the architecture of a prevention system, risk assessment, and modelling behaviour change.

• Victor Rențea, Andrei Vasilățeanu, Radu Ioanițescu, Luca Dan Șerbănați, Software Engineering-Inspired Approach to Prevention Healthcare, in Electronic Healthcare, Springer, vol. 91, 2012, pp. 109–113.
• Victor Rențea, Margareta Rențea, Mihai Isăroiu, Bogdan Niță, Radu Ioanițescu, Prevention Assistant - Risk Evaluation Based on Sparse Data, EIDWT 2012, pp. 158–165.
• Victor Rențea — two other conference papers from the doctoral period, concerning the architecture and experimentation of the prevention system.

Of these works, the first, which I also co-authored, establishes the initial conceptual framework of a prevention system inspired by software engineering and remains, in my view, one of the most representative contributions from the beginning of the research. The article on risk assessment from sparse data complements this direction well, even though the risk-assessment model in the thesis remained, in my opinion, insufficiently exploited.

Overall, the publications reflect the purpose of the doctorate well, even though the subject would have deserved still wider circulation in the scientific community. Victor possessed the computing knowledge needed to bring this research greater international visibility.

My contribution as supervisor

In Victor Rențea’s case, my role was important above all in formulating the topic and directing it toward an area in which computing would be used structurally rather than decoratively. The topic did not come from him; I was the one who attempted to connect his technological interests to a new and difficult field: disease prevention and health promotion.

I had to make him recover and understand a vast domain: e-health, EHRs, HL7 standardisation, prevention contexts, behaviour change, modelling the individual, and the role of intelligent systems. He had excellent computing preparation from the outset, but the thesis required more than technology: it required interdisciplinary synthesis and conceptual modelling.

I believe my principal contribution was precisely this attempt at synthesis: compelling him to leave the comfort of the purely technological and construct a coherent vision of a field in which computing, medicine, behavioural sciences, and even marketing inevitably meet. More concretely, I pushed him to connect what he knew very well— software architectures, modelling, and design—to the topics that had concerned me for some time: the EHR, the individual’s context, prevention, behaviour change, and the use of software agents in complex healthcare systems. Without this direction, the thesis might have remained a well-executed technological exercise; through it, the work acquired a broader meaning as interdisciplinary research.

A retrospective assessment

Viewed retrospectively, Victor Rențea’s thesis is one of the most ambitious among those I supervised in this area. It attempts to do several things simultaneously: model prevention, define its data and ontology, extend the EHR, construct a multi-agent system, and experimentally test its effects on behaviour change.

Naturally, such ambition has its costs. Not all parts of the thesis reach the same level of maturity. Some components, such as the virtual individual model or risk assessment from partial data, would have deserved greater development. Yet the breadth of the attempt remains one of its merits: Victor was not content with a point solution, but sought to give form to a general framework.

The experiment showed an average 58% increase in daily exercise time, while at the same time confirming the major difficulty of behaviour change: almost half of the participants remained at the same stage of change. This combination of a positive result and resistance to change seems to me one of the thesis’s most credible conclusions.

If I were to summarise the thesis’s specific purpose in a single formula, I would say that Victor attempted to unite four things that usually remain separate: preventive medicine, conceptual modelling of the relevant data, EHR infrastructure, and an intelligent agent-based software architecture. This ambition explains the work’s unevenness, but also gives it its real value. We are not dealing with a simple application, but with an attempt to formulate a general framework for computer-assisted, personalised prevention integrated into a broader medical ecosystem.

After the doctorate

After completing his doctorate, Victor devoted himself precisely to the technologies and topics that, to a certain extent, he had already encountered in my courses: J2EE, design patterns, software architectures, and later an even wider range of topics connected with clean system design, microservices, events, Spring, and AI-assisted code.

Today he is a highly visible trainer, with a strong personality and a presence in many professional environments. This path confirms retrospectively what was already apparent from the beginning: Victor was well prepared, curious, and capable of rapidly assimilating new technologies and paradigms. His doctorate may not have led to an academic career proper, but it undoubtedly represented an important stage in his intellectual and professional development.