Abstract
Digitalizing health data expands access to care but often outpaces the privacy safeguards that should accompany it. Türkiye ranks 27th on the 2024 UN E-Government Development Index, and its e-Nabız health record system covers 82% of the population. Yet the duties imposed by Article 6 of the Personal Data Protection Law (KVKK) and by the Regulation on Personal Health Data are structurally inadequate to protect data carrying a high risk of stigmatization, such as data on sexually transmitted infections (STIs), from breach: they are insufficiently specified, and enforced only once a disclosure has surfaced. This article argues that the resulting harm follows a self-reinforcing ‘stigmatization-spread cycle’, leaving the state’s positive obligations under Articles 17 and 20 of the Constitution unfulfilled. It develops this argument through three Constitutional Court judgments—Bülent Kaya, T.A.A. and Kâmil Selman Bilgin—read with eight Personal Data Protection Board decisions on unauthorized access, unlawful transfer and publication of health data. It further shows that the ‘necessity’ standard limiting the public health exception in Article 6(3)(e) is left without operative content, and identifies three techniques by which other legal orders have made that term reviewable. The article proposes classifying STI data as ‘very-high-risk’ and making ex ante impact assessments mandatory.
1. Introduction
This article argues that Türkiye’s welfare infrastructure contains a vulnerability that social welfare law has not addressed: the centralized digital platforms through which health services are delivered concentrate and expose the most sensitive data that patients generate while seeking care. The setting is constitutional. Article 2 declares Türkiye a social state (‘sosyal devlet’)1, and Article 56 gives that commitment content through the right to health, imposing on the state a duty to organize, finance and protect the delivery of healthcare—a mandate carried out programmatically through the 2003 Health Transformation Program and the Social Insurance and General Health Insurance Act No. 5510.2 Scholarship on data protection and welfare provision has mostly framed the tension between the state’s data-gathering powers and individual rights through welfare conditionality and fraud detection (Gantchev 2019), treating data protection as a limit imposed on the welfare state from outside. We argue the opposite: in a constitutional order built on the ‘sosyal devlet’ principle, protecting health data is part of the state’s own positive obligation under Article 56 to organize the delivery of healthcare. The benefits of the system are shared by the whole population; the risk of a confidentiality failure falls on its most exposed users. A welfare infrastructure that leaves those users open to harm which later remedies may not fully repair has not discharged that obligation, however extensive its technical coverage.
This vulnerability is not unique to Türkiye. In Türkiye, the legal framework has not kept pace with the scale and centralisation of the infrastructure. The resulting gap already exists. Türkiye’s rapid digital transformation in the health sector, characterized by its high ranking in the 2024 UN E-Government Development Index and near-universal health data coverage through the e-Nabız platform, makes this gap particularly apparent. The integration of digital health systems has dramatically expanded access to care—a genuine welfare achievement. However, the faster a state digitalizes its welfare infrastructure, the wider the gap that tends to open between what the system can do technically and what the law can guarantee in terms of rights protection. Digital welfare infrastructure can itself become a mechanism of harm rather than protection, as Australia’s Robodebt scheme showed.3
Türkiye’s experience presents a structural parallel: the same centralized digital infrastructure that delivers access to publicly financed healthcare—through the e-Nabız platform—simultaneously generates and exposes sensitive data in ways that the legal framework has not kept pace with. Harm is produced in this case by the structural inadequacy of privacy protection for data categories whose disclosure may lead to social exclusion that a later remedy may not fully repair. In both contexts, the state’s digital infrastructure outpaces its legal and protective safeguards, and the most vulnerable welfare beneficiaries bear the greatest cost. This article examines the processing of sexually transmitted infection (STI) data within Türkiye’s centralized digital health infrastructure. STI data is not merely sensitive in the ordinary sense. Writing about the criminalization of HIV transmission Weait (2019, pp. 594–96) uses ‘limit case’ for the rare instance of intentional transmission, which even committed decriminalization advocates find hardest to argue against. We borrow the term for a different purpose: STI data is the category whose disclosure tests most severely whether a data protection framework offers genuine rather than formal protection.
By analyzing Constitutional Court rulings and Personal Data Protection Board decisions, this study argues that the Act’s preventive duties are insufficiently specified for data of this kind and are enforced only once a disclosure has surfaced: the current ‘reactive model’ cannot prevent the social consequences of STI data disclosure, which a remedy granted afterwards may not fully repair. Consequently, a shift toward proactive, rights-based privacy legislation is needed to fulfil the state’s positive obligations under Articles 17 and 20 of the Constitution. All three decisions examined here—Bülent Kaya, T.A.A. and Kâmil Selman Bilgin—are individual applications (‘bireysel başvuru’), a mechanism introduced by the 2010 constitutional amendment4 and operative from 23 September 2012, under which a person whose fundamental rights have been violated by a public authority may apply directly to the Constitutional Court once ordinary remedies are exhausted. The jurisprudence they belong to is therefore young but fast-developing.
Since 2012, the Court has built a body of case law on positive obligations in the health domain. Under Article 17, which guarantees the right to protect and develop one’s physical and moral existence, those obligations are both substantive—an adequate regulatory framework for the protection of patients (İlker Başer and Others, §44)5—and procedural, requiring courts and administrative bodies to examine allegations of harm with the diligence Article 17 demands (Kâmil Selman Bilgin and Others, §35).6 They extend from physical harm to the integrity of health data: under Article 20, added in 2010, the Court has held that the positive obligation reaches conduct between private parties and requires, before all else, a legal framework that secures these rights effectively (T.A.A., §§76–77), and that adverse treatment in working life on the ground of health status engages Articles 17 and 20 together (§62), as Section 3 sets out.
Two institutions administer this field: the Ministry of Health, which operates the e-Nabız electronic health record platform and legislates for it by secondary regulation, and the Personal Data Protection Authority (‘Kişisel Verileri Koruma Kurumu’), whose Board issues binding decisions and administrative fines under Articles 15(5) and 18 of Law No. 6698.7 Both acquired their present digital remit during the same decade of institutional construction, producing the layered but incomplete governance structure this article examines.
Scholarship on Turkish data protection law has documented the KVKK’s general framework (Kınıkoğlu 2023) and sector-specific compliance gaps (Altun 2024); public health scholarship has documented HIV-related stigma in Türkiye (Gökengin et al. 2017); and Quinn (2019, p. 220) supplies the conceptual bridge, observing that privacy is bound up with stigmatization because people conceal what they fear will expose them to it. This framework operates at the level of information privacy theory, not at the constitutional and administrative intersection at issue here. What remains unexamined is whether the positive obligations under Articles 17 and 20 require a risk-differentiated regime for STI data. We read Bülent Kaya, T.A.A. and Kâmil Selman Bilgin together as one doctrinal trajectory—from the definition of personal data and the legality standard governing its processing, through the constitutional consequences of disclosure by a private actor, to the procedural duty to investigate systemic failures in healthcare provision. We connect that reading to the Personal Data Protection Board’s operational case law, which has not previously been read alongside the constitutional cases. And we propose a specific legislative mechanism rather than issuing a general call for reform. The trajectory is our synthesis, not a doctrine the Court has stated: none of the three decisions is on its own a health-data case in the full sense, and the claim is that read as a sequence they supply the definitional, substantive and procedural components of an obligation the Court has not yet articulated as one. The method is doctrinal, employing constitutional interpretation, case-law analysis, and policy analysis of the 2019 Regulation and its 2025 and 2026 amendments.
Türkiye is studied here as a case in its own right, and the article advances no comparative claim. It is chosen not because its digital infrastructure lags—Section 2 shows that it does not—but because that infrastructure is advanced enough to make the legal question a live one. Its rankings on the UN e-government indices, and the coverage of the platforms themselves, are given in Section 2. Two features support this choice: coverage is high enough that the processing is real rather than projected, and the individual-application jurisdiction opened in 2012 has produced a body of decisions in which the resulting failures are recorded. If a state that has digitized public health services at this scale meets structural limits in protecting a category such as STI data, those limits are not idiosyncratic.
2. Digital Pragmatism and Normative Gaps
The social state principle in Article 2 of the Constitution takes concrete form, in the health field, through the right to health in Article 56. The General Public Health Law No. 1593 of 19308 established the protection of public health as a state duty, including the mandatory reporting of infectious diseases. Built upon this foundation, the 2003 Health Transformation Program brought about a fundamental structural transformation, and universal health coverage was extended through the Social Insurance and General Health Insurance Act No. 5510: population coverage under social protection rose from 70% in 2000 to approximately 98% by 2010 (Akdağ 2012, p. 393).
Between 2008 and 2015, Türkiye established an integrated system encompassing various digital health platforms. Within this infrastructure, four systems have been implemented: e-Nabız, which consolidates a patient’s entire medical history in a single central repository and which had been adopted by 82% of the population by 2023 (Birinci 2023, p. 215); MEDULA, which manages billing and payment flows between hospitals and the state; the MHRS (Central Patient Appointment System), through which appointments are booked; and the e-prescription and e-report systems. This structure consolidates all data generated from primary care onwards into a single repository. As a result, Türkiye has surpassed countries such as Canada and the United Kingdom in terms of healthcare data coverage: on the proportion of key national health datasets covering 80% or more of the population, the figures are 73% for Türkiye, 60% for Canada and 28% for the United Kingdom, as reported by Timurtaş et al. (2024, p. 439), who source them in turn to OECD (2015, p. 31, Table 2.1) data. During the same period, the private healthcare sector also grew rapidly; the number of private hospitals rose from 271 in 2002 to a peak of approximately 572 in 2022, before declining to 552 by 2024, reaching approximately 35% of the total number of hospitals, whilst private sector bed capacity increased from 12,387 to 54,394, accounting for 20% of total capacity (T.C. Sağlık Bakanlığı [Turkish Ministry of Health] 2025, p. 130). According to the 2024 United Nations E-Government Survey data, Türkiye ranked 27th out of 193 countries on the E-Government Development Index (EGDI); whilst on the e-Participation Index (EPI), which measures digital participation, it ranked 22nd globally with a score of 0.7983 (United Nations 2024). Türkiye’s stance can be described as ‘digital pragmatism’, a governance model that prioritizes technical accessibility and administrative efficiency.
2.1. Defining ‘Digital Pragmatism’
‘Digital pragmatism’ is our own term. Turkish administrative law has developed less through a single codified procedural statute than through case law, secondary regulation (yönetmelik) and administrative circulars (genelge),9 a practice-driven mode that has given the administration considerable adaptive flexibility. We call it pragmatism because the orientation is to what works operationally: when that tradition meets rapid digitalization, public digital services are deployed and scaled with real speed, while the foundational legal questions—retention limits, access hierarchies, risk-differentiated categories, assessment before deployment—are settled incrementally alongside the system rather than before it. The term describes a governance style rather than philosophical pragmatism. This style helps explain both Türkiye’s digital health coverage and the lag in legal safeguards discussed below.
Indeed, Article 5(2) of the Regulation on Personal Health Data, which came into force in 2019,10 explicitly states that the system to be established aims to ‘deliver healthcare services effectively and swiftly’. The objective is legitimate. The risk is that speed and reach become the measures of success, and that the safeguards which would slow deployment are postponed.
Philip Alston, the UN Special Rapporteur on Extreme Poverty and Human Rights, describes the logic that drives the ‘digital welfare state’—efficiency, cost savings, fraud reduction, administrative modernization—and its systemic risk: not a neutral administrative exercise but a ‘revolutionary, politically driven’ restructuring in which citizens become ‘ever more visible to their governments, but not the other way around’, with rights-protective safeguards treated as secondary (Alston 2019, para. 3). Legality and transparency are among the casualties: major digital state systems have been implemented without adequate legal frameworks and without the prior risk assessments that would have identified structural vulnerabilities at the design stage rather than after harm (Alston 2019, para. 42–43). Türkiye’s data protection framework has likewise consolidated incrementally rather than by design.
Technology outpaces legal safeguards in what (Marchant 2011, pp. 22–26) calls the ‘pacing problem’: legislatures, regulators and courts slow as social complexity rises, while technology advances exponentially, so that technologies establish themselves in social life before a legal regime exists to govern them. This is Collingridge’s (1980, p. 19) dilemma of control: ‘attempting to control a technology is difficult, and not rarely impossible, because during its early stages, when it can be controlled, not enough can be known about its harmful social consequences to warrant controlling its development; but by the time these consequences are apparent, control has become costly and slow.’ Mass data processing platforms such as e-Nabız show the dilemma in operation.
2.2. STI Data and the Stigmatization-Spread Cycle
These shortcomings matter most for sexually transmitted infections, because the disclosure of STI data is not merely a records problem: its consequences—job loss, family breakdown, social exclusion, psychological harm—may be irreversible. A 2017 field study of 100 people living with HIV in Türkiye (the 2017 PLHIV Stigma Index) gives the local measure. Overall, 23.1% of participants reported having been directly subjected to stigma and discrimination and the overall rate of rights violations was 30%; gossip (69%) and verbal harassment or threats (46%) were the commonest social forms, while 30% had lost employment because of HIV-related stigma and 20% had been denied healthcare (Gökengin et al. 2017, pp. 15–16). Breach of confidentiality by health professionals was itself widespread: 43.9% were certain it had happened to them and a further 30.6% suspected that it had (p. 17).
McAllaster (2012, p. 8) records the same pattern, quoting a report of the ACLU AIDS Project: ‘[b]reaches of confidentiality can and do unravel people’s lives, forcing them to find new jobs, new schools, and new homes.’ A remedy granted afterwards may not reach much of that harm, which is why the leakage of STI data through digital health systems should be treated not as a technical security breach but as a violation of rights. In the adapted sense introduced above, STI data is a limit case: it tests the legal system where its protection is weakest.
The courts and the regulator have recorded both the failures and their consequences. The Personal Data Protection Board’s decisions on unauthorized access, unauthorized transfer and publication of health data record the failures themselves (Section 6), and can be read as an observational record of them. The Constitutional Court in T.A.A. records what follows from one: adverse treatment in working life once an employer knew an employee’s HIV status, in a judgement that sets out how stigma and exclusion operate on people living with HIV and how much more destructive their effects are in employment, where loss of work also jeopardizes access to lifelong treatment (§61) (Section 3). What they record together generates a self-perpetuating circle, which this article terms the ‘stigmatization-spread cycle’. Its mechanism proceeds in three stages. First, inadequate confidentiality protection in a digital or institutional setting produces or permits an unauthorized disclosure of STI status. Second, the disclosed individual suffers stigmatizing consequences—job loss, family breakdown, denial of care, social exclusion—which a later remedy may not fully repair. Third, awareness of that risk spreads among people living with or at risk of STIs and produces avoidance behaviour: non-disclosure, delayed testing, withdrawal from care. This last stage deepens the very coverage gap the digital system was built to close. Each stage feeds the next: the cycle reinforces itself, which is why a reactive, ex post legal model cannot arrest it.
Structuring stigma in this way has a precedent: Paterson et al.’s (2025) ‘hourglass’ model, built from 34 interviews across 25 outbreak-prone diseases, likewise traces a disclosure event through sequential domains into social harm. The cycle proposed here is our own formulation of that dynamic for STI data and Turkish digital health governance. The empirical support is substantial, though not uniform. Stangl et al. (2019, pp. 9–10) show that stigmatization disrupts care-seeking, participation and adherence, producing measurably lower testing rates and poorer continuity of HIV care. Correcting inflated perceptions of local stigma raised testing uptake among those who had overestimated it (Yu 2023, pp. 5, 8), though stigma-reduction interventions can misfire (Yang et al. 2023, pp. 14–16), and privacy concerns at facility and provider level are themselves a primary barrier to testing and care engagement (Tibbels et al. 2019). Withdrawal from care thus impairs the state’s own epidemiological knowledge as well as the individual’s health—a second reason why safeguards are needed before disclosure rather than remedies after it.
The cycle is not peculiar to Türkiye: the link between fear of privacy breaches and avoidance of treatment is documented globally (UNDP 2021; see also Yu 2023, p. 5; Tibbels et al. 2019). Ferguson et al.’s systematic review of HIV-stigma frameworks found legal and policy interventions as critical as individual-level ones, and the instruments for measuring structural stigma among the least developed in the literature (Ferguson et al. 2022). That gap applies to Türkiye, where the 2017 PLHIV Stigma Index (Gökengin et al. 2017) measures stigma as individuals experience it and no published instrument measures it at the level of institutions and legal design. We do not fill that empirical gap; we offer its legal-institutional counterpart—a record, in the Board’s decisions analyzed in Section 6, of how the design of the framework itself produces the conditions in which such harms recur.
3. Constitutional Duties to Protect Health Data
The protection of personal data under the Constitution of the Republic of Türkiye rests on Articles 17 and 20, and these provisions do not merely constitute a negative obligation prohibiting state intervention; they also impose a positive obligation requiring the state to establish active mechanisms to protect individuals’ data. That positive dimension is internal to the constitutional text rather than borrowed. Article 17(1) gives everyone the right to protect and improve their corporeal and spiritual existence, and Article 5 makes providing ‘the conditions required for the development’ of that same existence a fundamental duty of the state; the Turkish text of both provisions uses one phrase, ‘maddi ve manevi varlık’. The 1961 Constitution had already cast health as an express state duty. What the European Court of Human Rights and the United Nations treaty bodies supply is the apparatus for specifying the duty Articles 5 and 17 already impose, not its source.11 The third paragraph of Article 20, added by the 2010 constitutional amendment, opens by asserting that ‘everyone has the right to request the protection of personal data concerning them’. The same paragraph then sets the limits: personal data ‘may only be processed in cases provided for by law or with the explicit consent of the person’, and the principles and procedures of data protection ‘shall be regulated by law’. Those two sentences are the constitutional mandate for, and the constitutional ceiling on, the Personal Data Protection Law (KVKK) and the secondary legislation beneath it.
Bülent Kaya12 defines the scope of this framework. The application concerned police General Information Collection (GBT) records rather than health data, and the Court found no violation on the facts, by a majority over five dissents; the decision matters here for the definitional and legality principles it states, which later case law—including T.A.A., below—has applied to health data. The Court read personal data broadly: beyond name, surname and date of birth, any data permitting direct or indirect identification—telephone and passport numbers, CVs, photographs, images and audio recordings, fingerprints, genetic information, IP addresses, hobbies, preferences, group memberships and family details—falls within the category (§49). More importantly, it held that recording data relating to a person’s private life is itself an interference with the right to respect for private life under Article 20, whether or not the data is subsequently used (§53).
For STI data, the ruling’s ‘lawfulness’ requirement is particularly important. According to the Constitutional Court, the rules governing the processing of personal data cannot be satisfied merely by the existence of a formal statutory provision; these regulations must also be accessible, foreseeable and sufficiently precise (§§62–63). An individual must know in advance for how long their personal data will be stored, who has access to it, under what conditions it may be transferred, and how it will be destroyed; otherwise, the legality test cannot be passed (§64).
The classical human rights framework treats public health as a legitimate ground for limiting privacy: Article 8(2) ECHR permits interference where ‘necessary … for the protection of health’. Türkiye ratified the Convention in 1954 and accepted individual petition in 1987, so Strasbourg doctrine binds it, and the Constitutional Court’s legality test in Bülent Kaya (§§62–64) is the kind of proportionality safeguard the framework contemplates: a restriction is permissible only if it is accessible, foreseeable, and no broader than necessary. For most categories of data, this means a familiar structure, in which the individual’s interest in non-disclosure is weighed against, and may yield to, the collective interest in surveillance.
STI data resists this structure, because the two interests are not opposed but mutually constitutive. The public health case for collecting STI data rests on surveillance and treatment monitoring reducing transmission, yet inadequate confidentiality protection undermines that very objective: those who fear disclosure avoid testing, delay treatment and withdraw from the system whose effectiveness depends on their participation (Stangl et al. 2019, pp. 9–10; Yu 2023, p. 5). The European Court of Human Rights recognized this in Z v Finland,13 holding that ‘respecting the confidentiality of health data is a vital principle’ (§95), and that the point applies with particular force to HIV status, disclosure of which may expose a person ‘to opprobrium and the risk of ostracism’ and ‘discourage persons from seeking diagnosis or treatment and thus undermine any preventive efforts by the community to contain the pandemic’ (§96). That reading has been developed doctrinally (Taylor and Wilson 2019, p. 444) and supported empirically: inadequate confidentiality protection remains a central driver of HIV-related harm (Csete et al. 2023).
Weakening privacy protection to strengthen STI surveillance therefore undermines both objectives. In I v Finland, the Court held that an opportunity to claim damages was not enough: Article 8 requires ‘practical and effective protection to exclude any possibility of unauthorized access occurring in the first place’ (§47)14—the obligation of secure design that Section 6 applies to the Turkish context. Molldrem et al. (2025, p. 1) pose what they call ‘the fundamental ethical question’ for the US molecular HIV surveillance debate—whether the fact that clinical HIV data collected for the benefit of individual patients can also benefit population health means that they ought to be used for public health action—and the analysis here answers it in the negative for data of this kind.
This inversion carries a distributive dimension, and it is useful to separate the benefit from the means of delivering it. The benefit is access to publicly financed healthcare; the digital platforms described in Section 2 are the arrangements through which that benefit is now delivered. Their advantages—epidemiological visibility, system-wide efficiency, near-universal coverage—are diffuse and collective. The cost of a confidentiality failure—job loss, family breakdown, social exclusion—falls on the single individual whose data leaks, and may be irreversible in a way that an ex post sanction does not reach. The Board’s 2022/594 and 2023/767 decisions record disclosures of this kind and the fines that followed (Section 6). Patients who must weigh that risk before seeking care therefore enjoy the benefit on worse terms than other patients do. A framework that treats STI data as a routine public health exception under Article 6(3)(e), rather than as a very-high-risk category calling for safeguards before the fact, asks the most exposed participants in the welfare system to bear a risk whose advantages are shared by all. The state’s own expressive role compounds this: as Quinn (2019, p. 4) observes of stigmatizing state action generally, the law has difficulty regulating the stigmatizing effects of the state’s own communicative and administrative choices.
The cases central to this argument are disputes with private actors—a private employer in T.A.A., private hospitals and providers in the Board decisions of Section 6—yet they are adjudicated as individual applications under Articles 17 and 20. They therefore support not horizontal effect in the strict sense, that is, the direct application of constitutional rights between private parties, but the state’s duty to protect against private conduct: the obligation is unfulfilled because the regulatory framework failed to constrain a private actor. Legislative safeguards addressed to private data controllers, proposed in Section 7, are accordingly a coherent constitutional response.
T.A.A.15 is where these principles meet a concrete case. An HIV-positive worker told the physician at his workplace of his diagnosis, and the physician passed it to the employer with his consent; once the employer knew, he was kept away from the workplace on full pay for several months and then induced to resign. The Court held that he had been forced out on the ground of his health status (§88), and found a violation of Articles 17 and 20, read with the equality principle in Article 10. Two things follow, and they should be kept apart. The violation does not lie in the disclosure—the labour court rejected that complaint and the Constitutional Court did not disturb it—but in the reasoning of the domestic courts, which never examined the claim that the resignation was procured and never asked whether the employer was obliged to assess positions in which he could have worked without risk to others (§§90, 99). The Court accepted that protecting other employees was a legitimate aim, and held that the courts had nonetheless failed to balance it against the applicant’s rights (§§96–99). A second violation of Article 20 followed from the refusal, without adequate reasons, of his request that the hearings be closed to third parties (§112). That reasoning matters here for its structure: what the state owed was an assessment made before the applicant was excluded, not a remedy afterwards. The remedy he obtained was a retrial, ordered in 2017 for events of 2009, in proceedings the Court itself found unreasonably long (§116). The ruling thus brings the destructive stages of the stigmatization cycle within constitutional protection while showing the limits of a remedy that operates after the event. That a protective obligation was breached on these facts does not make every disclosure constitutionally impermissible, nor does it make the particular reforms proposed in Section 7 constitutionally required; the argument for them is ours, and it is that a framework which leaves such harms possible leaves the positive obligations under Articles 17 and 20 imperfectly discharged.
Kâmil Selman Bilgin and Others adds the right-to-life dimension of the same duty. A kidney-transplant patient was given a transfusion of Red Crescent blood infected with HIV, which the tests then in use could not detect, and died. The applicants’ complaint was that the Ministry of Health had failed to mandate the more advanced Nucleic Acid Testing (NAT) and to update the governing legislation. The Court did not rule on that complaint. What it found, after conflicting decisions in the administrative courts, was a procedural violation of the right to life: the administrative proceedings had left unanswered the critical questions raised by the administration’s own inspection reports, including why the advanced method was not in use (§§34–36). The distinction matters for the argument here. A procedural finding of this kind establishes that the state must account for its own technical choices where life and health are at stake; it does not itself hold that the choices were unlawful, and the inference we draw from it—that comparable duties of prior assessment should attach to the design of health data systems—is ours.
What is at stake in these cases is therefore not privacy narrowly conceived: because a person cannot decide about testing or treatment without knowing who will see the result, a confidentiality failure in STI data reaches the control over self-presentation that Article 17 protects as the ‘spiritual existence’ of the person, and it does so for those already exposed to stigma.16 Taken together, these rulings bear directly on the design of digital health systems. Full predictability for the individual—not least on retention periods for sensitive test results—is part of the legality standard Bülent Kaya states, and that standard is not met by administrative practice alone but requires clear and precise norms. What the Court has not said, and what we argue, is that for data of this kind the positive obligation under Article 20 is discharged only where protection is designed in before deployment rather than vindicated after it.
4. The Public Health Exception Under KVKK Article 6
The Personal Data Protection Law No. 6698 came into force in 2016, and two things brought it there. Article 20(3), added by the 2010 constitutional amendment, established for the first time an explicit constitutional right to the protection of personal data, which the European Commission recorded as a step toward Accession Partnership priorities on fundamental rights (European Commission 2010, pp. 80–84). The requirement then became a condition of accession policy: when the EU–Türkiye Visa Liberalization Dialogue was launched in December 2013, the Roadmap required Türkiye to enact a national data protection law and establish an independent supervisory authority as preconditions for visa-free travel (European Commission 2014, pp. 26–30). Law No. 6698 (KVKK), enacted in 2016, answered both at once: the requirement the Constitution had imposed in 2010, and the condition attached to visa liberalization (Kınıkoğlu 2023, pp. 26–27). That origin matters for what follows. A statute enacted to satisfy an accession condition is drafted to demonstrate the existence of a framework, which is not the same task as designing one around the risks particular categories of data carry; the gaps identified in this section are of a kind that the second task would have been likelier to catch.
Article 6(1) of Law No. 6698 classifies health and sexual life data as ‘special category personal data’, and Article 6(3) prohibits the processing of such data save on one of the grounds it lists. Article 6(2), which previously contained the prohibition, was repealed by Law No. 7499.17 The concept of special category or ‘sensitive’ data is, however, itself under-theorized in comparative data protection scholarship: Quinn and Malgieri (2021, pp. 1584, 1609–10) observe that the prevailing approach is already largely contextual and argue that it should attend more closely to the purpose of processing. Applying that analysis to KVKK Article 6(3), which subjects STI data to the same undifferentiated regime as other special category data whatever its stigmatization risk, suggests the risk tier proposed in Section 7—a conclusion we draw rather than one they advance.
Law No. 7499, adopted on 2 March 2024, ended the earlier dual regime. Under the former Article 6(3), special category data other than health and sexual life could be processed without consent wherever a statute so provided, whereas health and sexual life data could be processed without consent on one ground only, the grounds of health, and then only by persons under a duty of secrecy or by authorized institutions. The explanatory memorandum to the amending article records the pressure for change: under the former rule, only the Social Security Institution, the Ministry of Health and health institutions could process health data, yet such data was needed in the insurance sector, in labour legislation, in occupational health and safety, and in social services (reproduced in Kaya 2025, p. 109). Kaya adds the practical consequence, namely that the former structure confined the processing of special category data almost entirely to explicit consent (Kaya 2025, p. 8). Article 6(3) now lists eight grounds available to every special category, explicit consent first among them. Health and sexual life data accordingly gain the other seven too—notably Article 6(3)(f), on obligations in employment, occupational health and safety, social security and social services. The health ground survives as Article 6(3)(e), which permits processing ‘by persons under a duty of secrecy or by authorized institutions and organizations, where necessary for the protection of public health, preventive medicine, the carrying out of medical diagnosis, treatment and care services, and the planning, management and financing of health services’. Its wording is substantially that of the former provision (Altun 2024, p. 233), operating within what we would describe as a more systematic structure.
The reform separates a single unwieldy sentence into clauses and supplies a basis for processing in employment contexts through Article 6(3)(f). Structural problems nevertheless persist for data carrying a high risk of stigmatization. The phrase ‘protection of public health’ in Article 6(3)(e) retains its conceptual ambiguity, and the objective criteria by which the ‘necessity’ standard is to be assessed remain unspecified.
Other legal systems have given the same term reviewable content in three ways. A court can specify it: the Court of Justice holds that derogations from data protection must be limited to what is strictly necessary,18 and in Criminal Proceedings against V.S. it required that necessity for sensitive data be assessed ‘with particular rigour’ and be found ‘solely in a limited number of cases’ (§118),19 which in practice means asking whether the objective could have been achieved by processing only part of the data. A regulator can require it to be documented: under GDPR Article 35 large-scale processing of health data is presumptively high-risk and triggers a prior impact assessment in which necessity and proportionality must be assessed (Article 29 Data Protection Working Party 2017, pp. 9–10)20 Or a legislator can fix it as a rule, as HIPAA does by requiring covered entities to determine in advance which categories of staff need which categories of health information21 (U.S. Department of Health and Human Services 2003). None transfers directly—V.S. concerns law enforcement, and the HIPAA standard carries treatment exceptions—but together they show that the ambiguity in Article 6(3)(e) is a soluble drafting problem, and that the techniques for solving it are judicial, regulatory and legislative. The point bearing most directly on the Turkish framework is the European Data Protection Board’s—stated for law enforcement, but as a general criterion: necessity calls for objective criteria defining the circumstances in which processing may take place, so as to exclude ‘any processing of a general or systematic nature’ (European Data Protection Board 2023, §73)—which is what Article 5(2) of the Regulation examined in Section 5 authorizes.
The Act is not without preventive provisions. Article 6(4) requires that adequate measures determined by the Board be taken when special category data is processed, and Article 12(1) obliges controllers to take all necessary technical and organizational measures to prevent unlawful processing and unlawful access. The difficulty is twofold. The content of those duties is undefined, so what counts as an adequate measure for STI data is nowhere specified; they are enforced only once a disclosure has surfaced, since nothing requires a controller to demonstrate before deployment that it has assessed the risks the system creates. The Act has no equivalent of the impact assessment required by Article 35 of the GDPR, which mandates that ex ante analysis at the design stage. The 2024 amendment modernized the grounds for processing without addressing that gap.
Four structural issues follow. The first is the vagueness of ‘public health’, which leaves room for commercial transactions beyond treatment to fall within the exemption. That the insurance sector was among the intended beneficiaries is not a matter of inference: the explanatory memorandum names it first among the fields in which health data was needed. Article 6(3)(e) is confined to persons under a duty of secrecy and to authorized bodies, but the phrase ‘planning, management and financing of health services’ is not confined to treatment, and Article 6(3)(f) opens employment and social-security processing to health data directly. EIOPA (2021, p. 22) does not address STI data, but its account of how risk-pooling logics press insurers toward ever more granular health data illustrates where that pressure leads.
Turkish scholarship had identified the defect before 2024: Altundiş (2016, p. 318), reviewing the Strasbourg case law on medical data, concludes that a clear statutory rule must specify how, for what purpose and by what method such data may be collected, and that an administration cannot simply declare a measure to serve the public interest but must state how the processing and that purpose are connected. Article 6(3)(e) still imposes no such requirement. The safeguard it does impose is of a different kind: it restricts who may process—persons under a duty of secrecy, or authorized institutions—and requires that processing be necessary for the listed purposes. That is a safeguard attaching to the processor, not to the system. It says nothing about which categories of staff within an authorized institution may reach which records, for how long data is retained, whether access is logged, or whether the risks of a platform are assessed before it is built. In addition, ‘authorized institutions and organizations’ is itself left undefined. Article 9(2)(g) of the GDPR, by contrast, limits processing for ‘substantial public interest’ to measures proportionate to the aim and subject to safeguards laid down in Member State law. KVKK Article 4 states a general proportionality principle, but Article 6(3) attaches no comparable requirement of system-level safeguards to the public health ground.
The second issue is the absence of any hierarchy of risk among the special categories. Article 6(3) subjects data on race and ethnic origin to the same regime as data on health and sexual life, although their potential for stigmatization differs. The account of stigma in Section 2 supplies the reason this matters for the legal design: stigma is a structural exercise of power rather than a series of interpersonal acts (Parker and Aggleton 2003, pp. 17–19; Guise 2024, p. 2), it directly restricts access to healthcare (Stangl et al. 2019, pp. 1, 9), and Parker and Aggleton themselves conclude that it calls for structural and legislative intervention (pp. 22–24). An undifferentiated regime cannot register that difference, and so cannot calibrate safeguards to it.
The third issue is that the Act requires appropriate security without specifying what it requires at the design stage. Article 12 obliges controllers to take technical and organizational measures, but leaves it largely to their discretion to determine which; there is no counterpart to the duty of data protection by design and by default found in Article 25 of the GDPR. Cavoukian’s ‘Privacy by Design’ principles require privacy to be embedded in the architecture of information systems from the outset, and Article 25 of the GDPR has since made that a binding obligation (Cavoukian 2009, principles 1 and 3, p. 2). KVKK Article 12’s ‘appropriate level of security’, by contrast, tells a controller little about what to build. The European experience shows what such a formula can be made to mean: European Data Protection Board (2020, pp. 11–13, 25) translates the same open-ended wording into concrete expectations—data minimization by default, access restriction, storage limitation. It is that specification, rather than the abstract standard, which the third reform in Section 7 would import into Article 12.
The fourth and final issue concerns enforcement rather than the existence of duties. The obligations under Article 12 bind controllers from the outset, and under Article 15 the Board may open an investigation of its own motion. In practice, however, the Board’s attention is engaged by complaints and breach notifications, and the administrative fines of Article 18 presuppose a violation that has already been identified. For STI data, this sequencing matters more than it does elsewhere: the social harm begins at the moment of disclosure, and a sanction imposed afterwards may not repair it.
Decision 2022/594, discussed in Section 6, illustrates the gap at the operational level: the Board fined a controller for failing to take the technical and organizational measures Article 12(1) requires, but the disclosure of the employee’s test results to a third party at his workplace had already occurred, and the sanction could not undo it.
5. Safeguards Under the Personal Health Data Regulation
Secondary legislation on personal health data had a difficult passage. The first Regulation, issued by the Ministry of Health in October 2016, came six months after the KVKK entered into force, before the Board had been constituted and without its opinion. The Council of State stayed Articles 5(5) and 14 in July 2017, holding that the instrument did not conform to law because it had been made without the opinion the Board is required to give on draft legislation containing personal-data provisions, and at a time when the Board did not yet exist to determine the adequate measures Article 6(4) of the Act requires. In October 2018, it stayed in its entirety the instrument of 24 November 2017 that had amended the Regulation. The Regulation presently in force replaced that instrument in 201922; for a contemporaneous assessment, see Kaya (2019). The correction did not end there: the Council of State later annulled the provision of the 2019 Regulation that governed e-Nabız security settings, and the Ministry redrafted it only in December 2025.23 Secondary legislation was issued ahead of the institutional machinery meant to supervise it, and the courts corrected the result afterwards—the administrative counterpart of the reactive model the Board’s decisions display in Section 6.
The 2019 Regulation was enacted to operationalize Article 6(3) KVKK. It is not without protective provisions: Article 5(3) forbids compelling anyone to produce or display a printout of their past health data outside the processing grounds of Article 6(3) of the Act, and Article 5(4) and (5) require physical measures at counters and desks and the masking of printed test results. Three gaps in the Regulation’s safeguards leave it reliant on the penalty model.
The first issue is the breadth of the public health exception. Article 5(2) permits the Ministry to establish nationwide information systems to monitor everyone’s health status and deliver services more effectively, without naming the purposes for which the data may afterwards be used or the bodies that may consult it. Article 6 identifies five categories of physician who may access a record, but says nothing of the administrative, statistical and planning users of the national systems Article 5(2) authorizes. That is the uncertainty that matters: the patient cannot know, from the Regulation, who beyond those five categories will see the record. Read against the necessity standard discussed in Section 4, this is the provision on which most turns. A requirement of objective criteria that exclude ‘any processing of a general or systematic nature’ is not met by a power to monitor the health status of the whole population whose purposes are left unstated and whose users are left unnamed.
The second issue is the scope of access authorization and the weak application of the ‘need-to-know’ principle. Article 6/1 of the original Regulation allowed those responsible for providing healthcare services to access health data ‘provided that such access is limited to what is necessary for the provision of the healthcare service’, but it did not require the technical access controls that would enforce that limit at the design level—a gap whose consequences appear in the Board’s decisions examined in Section 6. The amendment of 3 December 202524 divided access authorization into five categories, each with its own access period: family physicians, consulting physicians, physicians employed by healthcare providers, physicians at the facility of admission, and emergency physicians. The exception is the family physician, whose access under Article 6(2)(a) carries no time limit at all. That is a real improvement, but it introduces no consent or notification mechanism, and phrases open to interpretation still leave the limits of access ambiguous: Article 6(5) permits the deactivation of security settings in cases of detention or imprisonment without any approval mechanism. The deeper limitation is that a time-limited window is a proxy for the need-to-know principle rather than an application of it. It bounds how long a category of physician may consult a record, but not whether the particular physician consulting it stands in any treatment relationship to the particular patient. That is the distinction on which the Board’s decisions in Section 6 turn, and which the third reform proposed in Section 7 would write into the Regulation.
The third issue is the absence of a required prior assessment. Breach notification under Article 18(2) and the sanctions to which Article 21 refers, in Articles 17 and 18 KVKK, necessarily operate after a violation—that is in the nature of notification and of penalties, and is not in itself an objection to them. The 2025 amendment indeed softened Article 21’s ‘their authorities shall be revoked’ to ‘the disciplinary authority shall take the necessary measures’. The precise shortcoming is the absence of any required prior assessment: nothing in the 2019 Regulation or its 2025 and 2026 amendments corresponds to the impact assessment that GDPR Article 35 requires before a system processing sensitive data on a large scale is deployed. The missing safeguard would apply before deployment.
On 4 July 2026, the Ministry added Article 13/A, under which a person may ask for a diagnosis already on the record to be re-evaluated against their current health status, and which provides that past diagnoses may not be relied on in any process, recruitment above all.25 The provision is valuable, and it addresses a harm this article documents: in T.A.A., the applicant’s HIV status, once known to his employer, cost him his job. But it operates on the record after the entry exists and after the disclosure risk has been run; it governs what may be done with a diagnosis, not who may see it or on what conditions a system exposing it may be built. Where that condition should sit is the subject of Section 7: a new Article 6-a of this Regulation, modelled on Article 35 of the GDPR, binding the controller before the system goes live rather than the Board after it has failed.
6. Health Data Breaches in the Board’s Decisions
The Board’s health-data decisions document the limits of the reactive model. They cluster around four recurring failures—control over who may access a record, control over onward transfer, control over publication, and the absence of any mechanism operating before disclosure—each of which bears directly on the structural risks associated with STI data.
The Board’s Decision No. 2021/962 confirms a control gap in the technical design of a digital data platform. A data subject discovered that a physician’s secretary—with whom the data subject had no patient–physician relationship—had accessed the data subject’s information on the digital platform using the physician’s system password. The Board ruled that the hospital, as the data controller, had failed to implement reasonable technical and administrative measures to prevent unlawful access to personal data under Article 12/1-b of the KVKK.26 The systemic implication of the decision extends beyond the penalty imposed on a single hospital. The technical design of the platform leaves it open to the practical delegation of a physician’s access privileges to third parties—a secretary, in this instance—and to their exercise without passing through any filter of an active treatment relationship. Decision No. 2023/695 records the same failure at a private medical centre, where uncertainty about access periods produced unauthorized access.27
Decision No. 2021/761 states the design problem most explicitly. A pediatrician at a public training and research hospital retrieved a child’s psychiatric records—outside his own specialty and without any treatment relationship in that field—and released them to the opposing party’s lawyer in a custody dispute. The Board identified two features of the hospital’s system. First, every doctor and every polyclinic registration clerk in the hospital could view every patient’s file; the Ministry of Health defended this as inherent to clinical practice, and the Board rejected the defence, holding the arrangement contrary to the principle that processing be connected, limited and proportionate to its purpose, and ordering the controller to construct an access-authorization matrix. Second, the system logged additions, deletions and modifications but did not log viewing at all, so it was impossible to establish who had looked at which record. The Board ordered the system updated to log access, including mere viewing.28
The second finding is the Turkish counterpart of I v Finland, where a records system that could not identify who had accessed a file was itself held to breach the obligation of secure design. The Board applied that standard and ordered remedies, but could do so only after a child’s psychiatric records had reached a custody file. Because the controller was a public institution, no administrative fine was available: fines only apply to private controllers, while for public institutions, Article 18(3) (now Article 18(4)) KVKK substitutes a disciplinary route. The Board could therefore issue only instructions and a disciplinary referral.
This pattern—access-control failure enabling unauthorized viewing of sensitive records by staff without a treatment relationship to the patient—is not unique to Türkiye: Lenton et al. (2025, p. 5) document, from interviews with 30 clinical and non-clinical healthcare workers in Australia, how ‘control over who has access to records containing “sensitive” information can be undermined and breached’ by the way health-record systems are designed and used, reproducing structural stigma through the system’s own affordances rather than through any single actor’s misconduct. They further record (p. 6) that efforts to restrict access according to a person’s role or the relevance of the information were ‘undermined by insufficiently sophisticated systems’. The Board’s 2021/962 and 2023/695 decisions are, in this sense, a Turkish instance of a design problem documented independently in comparative literature.
A second channel in the Board’s case law has no counterpart in the constitutional cases: the community pharmacy. Under the Medula Eczane module of the Social Security Institution’s billing system, every contracted pharmacy can retrieve, on a national identity number alone, a person’s prescriptions, and medical reports for the preceding twelve months. Three decisions show what that permits. In 2020/35529 a pharmacist’s spouse—not a health worker at all—obtained a competitor’s diagnosis and medication printouts and attached them to a petition seeking to block the competitor from opening premises. In 2023/113030, the pharmacist printed reports and medication records from Medula and handed them to the data subject’s former spouse for use in a custody dispute. The Board’s earliest published health-data decision, 2018/143,31 concerns the same setting, a pharmacy disclosing to a third party the health data of a person on prescription medication.
Three features matter. The failure recurs from 2018 to 2023, which is hard to explain as individual carelessness. Access is keyed to an identity number rather than to a treatment relationship, so those who can view a person’s medication history are not the clinicians treating them but every contracted pharmacy in the country. Medication records disclose diagnosis by inference: an antiretroviral regimen identifies HIV status as effectively as a test result. The Board has thus documented, three times, a disclosure route to which STI data is fully exposed, without any of those cases being an STI case.
In 2023/1130, the Board defined pharmacist confidentiality by reference to information whose disclosure would cause the patient to be shamed and excluded.32 The language here is that of stigma rather than of data security, and it suggests the framework applied here is already latent in the Board’s reasoning.
Decision No. 2022/594 adds the transfer channel. Drug test results were emailed to an unauthorized person at the data subject’s workplace without explicit consent; the Board found no legal basis under Article 6 and imposed an administrative fine.33 The implication for STI data is direct: if biological test results can pass from a healthcare facility to an employer unchecked, the same route is open to data whose disclosure carries a far higher risk of stigmatization.
The Board’s Decision No. 2023/767 is not an STI case—consistently with the search reported below, which found none—but it shows in the starkest form the kind of social harm that disclosure of a stigmatizing diagnosis can produce. Sensitive health data belonging to a married couple receiving treatment at a private hospital was published in a newspaper article using their full names. The Board ruled that security obligations under Article 12/1 had been violated and imposed an administrative fine of 100,000 TL (approx. 2230 US Dollars).34 Removal of the article after publication may not fully repair the consequences of disclosure. The sequence shows why the reactive model falls short: by the time administrative intervention takes effect, the disclosure has already been made.
The common thread is that the Board intervenes only after social and legal harm has occurred. Nor is that reactivity peculiar to the KVKK regime: Turkish administrative-law doctrine has long treated the administration’s failure to retain and protect medical data—and a fortiori its disclosure to third parties without consent—as a service fault (hizmet kusuru) giving rise to liability in damages (Altundiş 2016, pp. 344–45). That route shares the same limitation as the Board’s fining power: it compensates after the fact, and cannot restore the confidentiality whose loss is the harm.
A keyword search of the Board’s published decision summaries returned 67 decisions, and not one concerns HIV status or a sexually transmitted infection (Table S1).35 We do not treat that absence as proof of its own cause: published decisions may under-represent STI-related breaches where stigma discourages complaints, and the Board may in any event act of its own motion under Article 15, so the record is not a census. What the distribution does show is that health-data breaches reach the Board—from pharmacies, hospitals, employers and newspapers—and that STI-data breaches do not appear among them.
The mechanism is the enforcement counterpart of the cycle set out in Section 2. A complaint to the Board requires the applicant to be named and the condition to be recorded in the file. For someone whose STI status has just been disclosed without consent, and who is at elevated risk of dismissal and refusal of care, seeking a remedy therefore repeats in an official register the exposure that caused the injury—a calculation the 2017 PLHIV Stigma Index shows to be rational, with 30% of respondents having already lost employment and 20% having been refused healthcare (Gökengin et al. 2017, pp. 15–16).
The result is a selection effect: the docket records the breaches victims were willing to be named in, and is unlikely to record those they cannot afford to be named in. This remains a hypothesis: the absence of STI cases from the published summaries is consistent with it but does not itself tell us why they are absent.
Two consequences follow.36 For STI data, the reactive model is not merely slow; in a predictable subset of the most serious cases, it is never triggered, because the remedy is conditioned on a disclosure the victim has every reason to refuse. The eight decisions analyzed above should be read as the visible portion of a larger distribution rather than as its measure: the argument rests not on their number but on their being structurally representative, each documenting a design failure—of access control, transfer control or publication control—that operates independently of the data category and would operate identically, with graver consequences, on STI data. This is why an obligation binding before deployment is the appropriate remedy: it does not depend on a victim coming forward.
7. Conclusions and Recommendations
The Board’s decisions show delayed and reactive repair rather than a legal system incapable of repair. Because the existing measures are enforced only once harm has come to light, they cannot prevent the social consequences of disclosing data such as STI status, and a remedy granted afterwards may not fully repair them. The benefit at stake is access to publicly financed healthcare, increasingly organized through digital systems, and the question is whether patients with stigmatized diagnoses can obtain it without bearing additional and inadequately controlled risks of disclosure. We argue that extending healthcare coverage does not by itself discharge the state’s protective obligations where confidentiality risks remain inadequately controlled.
We propose three reforms, each specified by the provision it would amend, the body responsible for oversight and its enforcement mechanism.
First, a ‘very-high-risk data’ sub-category should be added to KVKK Article 6, most naturally as a new Article 6(3-a), defining STI and HIV status data as data whose processing triggers the safeguards below in addition to the existing Article 6(3) grounds. The lawful grounds are not withdrawn, and the safeguards are context-dependent in their stringency: they attach most heavily to uses that extend access beyond an active treatment relationship—population-level monitoring, billing and insurance, employment-related processing—and most lightly to treatment by the clinician treating the patient. This adds a risk tier to the 2024 reform’s unified structure rather than abandoning it. Quinn and Malgieri (2021, pp. 1609–10) make the case for closer attention to the purpose of processing; the tier is how we would give that effect in Article 6. Oversight of the classification and of data controllers’ compliance with it should sit with the Personal Data Protection Board (KVK Kurulu), which already exercises this function for Article 6 generally; enforcement would proceed through the existing Article 18 administrative fine mechanism, but with a statutory aggravating factor for very-high-risk data breaches, so that fines for STI data leaks are set at a materially higher tier than the 75,000–100,000 TL range seen in the 2022/594 and 2023/767 decisions discussed in Section 6.
Second, a mandatory Data Protection Impact Assessment (DPIA) requirement should be introduced into the Regulation on Personal Health Data, most naturally as a new Article 6-a modelled on GDPR Article 35, requiring any data controller introducing or materially modifying a digital system that processes very-high-risk data to conduct and document a DPIA before the system is deployed, addressing access controls, retention limits, and breach-response design. What we propose is a duty to assess and document, coupled with a duty to submit the assessment to the Board, not prior approval. Deployment would not await the Board’s consent, but the Board would hold the assessment and could act before a breach, as under the GDPR’s prior-consultation mechanism in Article 36 rather than any authorization procedure. The parallel lies only in the Board’s ability to act before a breach without an authorization procedure: Article 36 requires consultation only where the assessment indicates a high risk that would remain without further mitigation, whereas the duty proposed here requires submission for every such system, whatever the assessment finds. Enforcement should attach non-compliance to the Article 18 fining power as an independent violation, so that deploying such a system without a completed assessment is sanctionable whether or not a breach has occurred. For public controllers the position needs care, because the Act already provides a route. Administrative fines under Article 18(1) are applied, by Article 18(2), to controllers who are natural persons or private-law legal persons; where the listed acts occur within a public institution, Article 18(4) substitutes a different mechanism, under which the Board notifies the institution, disciplinary proceedings are taken against the officials concerned and the result must be reported back to the Board. What that mechanism reaches, however, is the individual official rather than the institution, and conduct rather than design. In 2021/761, the Board could order the hospital to build an access-authorization matrix and to log viewing, but no consequence attaches to the institution itself if such an order goes unimplemented. The gap to be closed is therefore narrower than it first appears: the assessment duty should identify an officer accountable for it, so that the Article 18(4) route reaches the design failure and not only the member of staff who looked.
Third, an enforceable privacy-by-design obligation should be added to KVKK Article 12, specifying—where Article 12’s current ‘appropriate level of security’ standard is silent—minimum technical requirements for very-high-risk data: granular, need-to-know access controls tied to an active treatment relationship (addressing the access-control failures documented in the Board’s 2021/962 and 2023/695 decisions and in Lenton et al.’s (2025, pp. 5–6) comparable findings), immutable access logging, whose absence 2021/761 exposed, and default retention limits absent an explicit extension. Oversight should combine the KVK Kurulu’s existing audit powers with a technical-standards role for the Ministry of Health, which already issues the secondary regulations governing e-Nabız access tiers; enforcement would proceed through the Article 18 penalties for breach of Article 12. For repeated non-compliance, we would add a consequence the Act does not presently provide. Article 15(7) lets the Board halt processing only where manifest unlawfulness coincides with harm difficult or impossible to repair;37 what is missing is a category-scoped response to persistent technical non-compliance. The Board should be able to suspend a controller’s authorization to process very-high-risk data specifically, leaving its other processing—and the patient care depending on it—untouched, until the required technical measures are in place. The direction of travel has lately been the other way: until December 2025, Article 21(2) of the Regulation provided that the authorizations of officials who failed to meet its requirements would be revoked, and the amendment replaced that with a referral to the disciplinary authority. We would restore a consequence of that kind but attached to the controller’s processing of the category at issue rather than to an individual’s authorizations, and placed in the Act rather than in secondary legislation.
First, let us consider oversight capacity: the KVK Kurulu is a comparatively young institution, and assigning it DPIA review and technical-standards functions in addition to its existing complaint-adjudication role will require a corresponding expansion of its technical staffing—a resourcing question beyond the scope of this article but a precondition for the reforms proposed here to function as intended. Second, let us consider consent design: the opt-out affordances, dynamic consent and plain-language notification of access identified by Molldrem et al. (2023, p. 1036) offer a model for granular patient control over which categories of provider may access very-high-risk data, and secondary legislation implementing the third proposal should prefer such mechanisms to a binary access/no-access design. Molldrem and Smith’s (2020, pp. 19–20) ‘HIV data justice’ supplies the accompanying caution: consent alone cannot answer structural harms, so consent design supplements the design obligations proposed above rather than substituting for them.
When digital health infrastructure develops without matching legal safeguards, the most exposed patients bear the resulting risks, as Türkiye’s experience illustrates. The urgent structural step is to define HIV and STI data as a very-high-risk category in the governing legislation, so that they are shared only where a lawful ground genuinely requires it and subject to the safeguards set out above. Protecting these data is not only a matter of individual privacy. Confidentiality is what sustains confidence that a diagnosis can be sought without cost, and so what keeps the door to care open.
Supplementary Materials
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/laws15060167/s1, File S1: Search log of the Personal Data Protection Board’s published decision summaries (keyword ‘sağlık’, searched 3 September 2026).
Author Contributions
Conceptualization, B.B.S.; methodology, Ö.A.; investigation, B.B.S.; resources, B.B.S. and Ö.A.; writing—original draft preparation, B.B.S. and Ö.A.; writing—review and editing, B.B.S. and Ö.A.; project administration, B.B.S. and Ö.A. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were generated by experiment or measurement. The Constitutional Court judgments, Personal Data Protection Board decisions and legislative materials analyzed are publicly available through the sources cited. The one dataset compiled for this article is the keyword search of the Board’s published decision summaries reported in Section 6 (keyword ‘sağlık’, 67 results, decisions dated 5 December 2018 to 26 December 2024, searched 3 September 2026). The search log, listing each decision number, date, subject-matter classification and permanent address, is supplied as supplementary material (Table S1) and is also available from the corresponding author.
Acknowledgments
During the preparation of this manuscript, the authors used Gemini 1.5 Flash and Claude (Sonnet 5.5) for three purposes: translating the authors’ Turkish-language drafts into English; refining the linguistic structure of the English text for academic clarity; and, during this revision, cross-checking citations, quotations and pinpoint references against their original sources. The research question, the theoretical framework, the argument, the selection and interpretation of all legal and scholarly sources, and all substantive conclusions are the authors’ own. No text was generated by these tools without the authors’ direction, and no source was cited on the basis of a tool’s output alone. The authors have verified every citation in this revised manuscript against the source itself, have reviewed and edited all output, and take full responsibility for the content of this publication.
Conflicts of Interest
The authors declare no conflicts of interest.
Abbreviations
| DPIA | Data Protection Impact Assessment |
| ECHR | European Convention on Human Rights |
| EGDI | E-Government Development Index |
| EIOPA | European Insurance and Occupational Pensions Authority |
| EPI | E-Participation Index |
| GDPR | General Data Protection Regulation |
| HIPAA | Health Insurance Portability and Accountability Act |
| KVK Kurulu | Kişisel Verileri Koruma Kurulu (Personal Data Protection Board) |
| KVKK | Kişisel Verilerin Korunması Kanunu (Personal Data Protection Law) |
| MHRS | Merkezi Hekim Randevu Sistemi (Central Patient Appointment System) |
| NAT | Nucleic Acid Testing |
| PLHIV | People Living with HIV |
| STI | Sexually Transmitted Infection(s) |
| UN | United Nations |
| UNDP | United Nations Development Programme |
References
- Akdağ, Recep, ed. 2012. Türkiye Sağlıkta Dönüşüm Programı Değerlendirme Raporu (2003–2011). Ankara: T.C. Sağlık Bakanlığı. [Google Scholar]
- Alston, Philip. 2019. Digital Welfare States and Human Rights: Report of the Special Rapporteur on Extreme Poverty and Human Rights. UN Doc A/74/493. Available online: https://documents.un.org/doc/undoc/gen/n19/312/13/pdf/n1931213.pdf (accessed on 25 July 2026).
- Altun, Abdurrahim. 2024. Tıp Hukuku Çerçevesinde Kişisel Sağlık Verilerinin İhlalinden Doğan Hukukî Sorumluluk. Hacettepe Hukuk Fakültesi Dergisi 14: 219–53. [Google Scholar] [CrossRef] [Scilit]
- Altundiş, Mehmet. 2016. Tıbbi Kişisel Verilerin Tutulması ve Korunması Yükümlülüğü ve İdarenin Bu Yükümlülüğünü Yerine Getirmemesinden Doğan Sorumluluğu. Türkiye Adalet Akademisi Dergisi 7: 313–51. Available online: https://dergipark.org.tr/tr/pub/taad/article/694118 (accessed on 25 July 2026).
- Article 29 Data Protection Working Party. 2017. Guidelines on Data Protection Impact Assessment (DPIA) and Determining Whether Processing Is ‘Likely to Result in a High Risk’ for the Purposes of Regulation 2016/679. WP248 rev.01. Adopted 4 April 2017, Last Revised and Adopted 4 October 2017; Endorsed by the European Data Protection Board on 25 May 2018. Available online: https://ec.europa.eu/newsroom/article29/items/611236 (accessed on 3 September 2026).
- Birinci, Şuayıp. 2023. A Digital Opportunity for Patients to Manage Their Health: Turkey National Personal Health Record System (The e-Nabız). Balkan Medical Journal 40: 215–21. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cavoukian, Ann. 2009. Privacy by Design. Toronto: Information and Privacy Commissioner of Ontario. Available online: https://www.ipc.on.ca/en/media/1826/download (accessed on 25 July 2026).
- Collingridge, David. 1980. The Social Control of Technology. London: Frances Pinter. [Google Scholar]
- Commonwealth of Australia. 2023. Effects of Robodebt on Individuals. In Royal Commission into the Robodebt Scheme, Report. Brisbane: Royal Commission into the Robodebt Scheme, vol. 2, chap. 10. Available online: https://robodebt.royalcommission.gov.au/publications/report (accessed on 25 July 2026).
- Csete, Joanne, Richard Elliott, and Edwin J. Bernard. 2023. So Many Harms, So Little Benefit: A Global Review of the History and Harms of HIV Criminalisation. The Lancet HIV 10: e52–61. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- EIOPA. 2021. Artificial Intelligence Governance Principles: Towards Ethical and Trustworthy Artificial Intelligence in the European Insurance Sector. Frankfurt: European Insurance and Occupational Pensions Authority. Available online: https://www.eiopa.europa.eu/system/files/2021-06/eiopa-ai-governance-principles-june-2021.pdf (accessed on 25 July 2026).
- European Commission. 2010. Turkey 2010 Progress Report. SEC(2010) 1327 final. Brussels: European Commission. [Google Scholar]
- European Commission. 2014. Report on Progress by Turkey in Fulfilling the Requirements of Its Visa Liberalisation Roadmap. COM(2014) 646 final. Brussels: European Commission. [Google Scholar]
- European Data Protection Board. 2020. Guidelines 4/2019 on Article 25 Data Protection by Design and by Default. Version 2.0. Brussels: European Data Protection Board. [Google Scholar]
- European Data Protection Board. 2023. Guidelines 05/2022 on the Use of Facial Recognition Technology in the Area of Law Enforcement. Version 2.0. Adopted 26 April 2023. Brussels: European Data Protection Board. [Google Scholar]
- Ferguson, Laura, Sofia Gruskin, Maria Bolshakova, Sachi Yagyu, Ning Fu, Nicole Cabrera, Mary Rozelle, Kevin Kasoka, Tessa Oraro-Lawrence, Lucy Stackpool-Moore, and et al. 2022. Frameworks and Measures for HIV-Related Internalized Stigma, Stigma and Discrimination in Healthcare and in Laws and Policies: A Systematic Review. Journal of the International AIDS Society 25: e25915. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Gantchev, Valentin. 2019. Data Protection in the Age of Welfare Conditionality: Respect for Basic Rights or a Race to the Bottom? European Journal of Social Security 21: 3–22. [Google Scholar] [CrossRef] [Scilit]
- Gökengin, Deniz, Şebnem Çalık, and Pınar Öktem. 2017. Türkiye’de HIV’le İlgili Damgalama ve Ayrımcılığın Analizi: HIV’le Yaşayan Kişiler İçin Damgalanma Göstergesi Sonuçları. Klimik Dergisi 30: 15–21. [Google Scholar]
- Guise, Andy. 2024. Stigma Power in Practice: Exploring the Contribution of Bourdieu’s Theory to Stigma, Discrimination and Health Research. Social Science & Medicine 347: 116774. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kaya, Mehmet Bedii. 2019. Sağlık Bakanlığı’nın Kişisel Sağlık Verileri Hakkında Yönetmeliğine İlişkin Değerlendirmeler. Available online: https://mbkaya.com/saglik-bakanligi-kisisel-saglik-verileri-yonetmeligi/ (accessed on 25 July 2026).
- Kaya, Mehmet Bedii. 2025. KVKK Reformu: 2024 Değişiklikleri, Dijital Baskı 2.0. Available online: https://mbkaya.com/hukuk/kvkk-reformu.pdf (accessed on 2 October 2026).
- Kaya, Mehmet Bedii, and Furkan Güven Taştan. 2025. Kişisel Veri Koruma Hukuku: Mevzuat—İçtihat—Bibliyografya. Çevrimiçi Sürüm 3.0, Current as at January 2025. Available online: https://mbkaya.com (accessed on 2 October 2026).
- Kınıkoğlu, Batu. 2023. Implementing a New Data Protection Law: Lessons from the Turkish Experience. International Data Privacy Law 13: 25–46. [Google Scholar] [CrossRef] [Scilit]
- Lenton, Emily, Kate Seear, Adrian Farrugia, Chris Lemoh, Elena Cama, Gemma Nourse, and Carla Treloar. 2025. Health-Related Stigma: The Affordances of Electronic Health Management Systems in the Production of Structural Stigma. Sociology of Health & Illness 47: e70043. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lupton, Deborah. 2014. Critical Perspectives on Digital Health Technologies. Sociology Compass 8: 1344–59. [Google Scholar] [CrossRef] [Scilit]
- Marchant, Gary E. 2011. The Growing Gap between Emerging Technologies and the Law. In The Growing Gap Between Emerging Technologies and Legal-Ethical Oversight: The Pacing Problem. Edited by Gary E. Marchant, Braden R. Allenby and Joseph R. Herkert. Dordrecht: Springer, pp. 19–33. [Google Scholar]
- McAllaster, Carolyn. 2012. HIV-Related Stigma: Protecting the Confidentiality of Clients Living with HIV/AIDS. Durham: Center for HIV Law and Policy Resource Bank. Available online: https://www.hivlawandpolicy.org/sites/default/files/HIV-Related%20Stigma%20-%20Protecting%20the%20Confidentiality%20of%20Clients%20Living%20with%20HIV%20AIDS%20(McAllaster).pdf (accessed on 25 July 2026).
- Molldrem, Stephen, and Anthony K. J. Smith. 2020. Reassessing the Ethics of Molecular HIV Surveillance in the Era of Cluster Detection and Response: Toward HIV Data Justice. The American Journal of Bioethics 20: 10–23. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Molldrem, Stephen, Anthony K. J. Smith, and Alexander McClelland. 2023. Advancing Dialogue about Consent and Molecular HIV Surveillance in the United States: Four Proposals Following a Federal Advisory Panel’s Call for Major Reforms. The Milbank Quarterly 101: 1033–46. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Molldrem, Stephen, Anthony K. J. Smith, Cecilia Chung, and Alexander McClelland. 2025. HIV Data and Public Health Ethics. Public Health Ethics 18: phaf008. [Google Scholar] [CrossRef] [Scilit]
- Naarttijärvi, Markus. 2023. AI and Sensitive Personal Data under the Law Enforcement Directive: Between Operational Efficiency and Legal Necessity. In YSEC Yearbook of Socio-Economic Constitutions 2023. Cham: Springer, pp. 331–57. [Google Scholar] [CrossRef] [Scilit]
- OECD (Organisation for Economic Co-operation and Development). 2015. Health Data Governance: Privacy, Monitoring and Research. OECD Health Policy Studies. Paris: OECD Publishing. [Google Scholar] [CrossRef] [Scilit]
- Parker, Richard, and Peter Aggleton. 2003. HIV and AIDS-Related Stigma and Discrimination: A Conceptual Framework and Implications for Action. Social Science & Medicine 57: 13–24. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Paterson, Amy, Benjamin Jones, Olive Kabajaasi, Ashleigh Cheyne, Harun Tulunay, Kkunsa Hadson, Jeni Stolow, Nina Gobat, Piero Olliaro, and Amanda Rojek. 2025. An Hourglass Model for Conceptualising Stigma in Infectious Disease Outbreaks. Scientific Reports 15: 15339. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Quinn, Paul. 2019. Stigma, State Expressions and the Law: Implications of Freedom of Speech. Abingdon: Routledge. [Google Scholar]
- Quinn, Paul, and Gianclaudio Malgieri. 2021. The Difficulty of Defining Sensitive Data—The Concept of Sensitive Data in the EU Data Protection Framework. German Law Journal 22: 1583–612. [Google Scholar] [CrossRef] [Scilit]
- Rodin, Siniša. 2025. Proportionality Analysis in ‘Digital’ Cases. In Remedies to Digital Vulnerability in European Private Law. Edited by Camilla Crea and Marta Infantino. Cham: Springer, pp. 171–85. [Google Scholar] [CrossRef] [Scilit]
- Stangl, Anne L., Valerie A. Earnshaw, Carmen H. Logie, Wim van Brakel, Leickness C. Simbayi, Iman Barré, and John F. Dovidio. 2019. The Health Stigma and Discrimination Framework: A Global, Crosscutting Framework to Inform Research, Intervention Development, and Policy on Health-Related Stigmas. BMC Medicine 17: 31. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Taylor, Mark J., and James Wilson. 2019. Reasonable Expectations of Privacy and Disclosure of Health Data. Medical Law Review 27: 432–60. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- T.C. Sağlık Bakanlığı [Turkish Ministry of Health]. 2025. Sağlık İstatistikleri Yıllığı 2024; Ankara: Sağlık Bilgi Sistemleri Genel Müdürlüğü. Available online: https://www.saglik.gov.tr/TR-114952/saglik-istatistikleri-yilligi-2024.html (accessed on 25 July 2026).
- Tibbels, Natalie J., Zoé M. Hendrickson, Danielle A. Naugle, Abdul Dosso, Lynn M. Van Lith, Elizabeth C. Mallalieu, Anne Marie Kouadio, Walter Kra, Diarra Kamara, Patricia Dailly-Ajavon, and et al. 2019. Men’s Perceptions of HIV Care Engagement at the Facility- and Provider-Levels: Experiences in Côte d’Ivoire. PLoS ONE 14: e0211385. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Timurtaş, Meral, Tutku Ekiz Kavukoğlu, and Mehveş Tarım. 2024. Sağlık Verilerinin Gizliliği Güvenliği ve Erişilebilirliğine Dair Yasal Düzenlemeler: Kanada İngiltere ve Türkiye Karşılaştırması. Toplum ve Hekim 39: 435–46. [Google Scholar]
- UNDP. 2021. Guidance on the Rights-Based and Ethical Use of Digital Technologies in HIV and Health Programmes. New York: United Nations Development Programme. Available online: https://www.undp.org/publications/guidance-rights-based-and-ethical-use-digital-technologies-hiv-and-health-programmes (accessed on 25 July 2026).
- United Nations. 2024. E-Government Survey 2024. New York: UN DESA. Available online: https://publicadministration.un.org/egovkb (accessed on 25 July 2026).
- U.S. Department of Health and Human Services, Office for Civil Rights. 2003. Minimum Necessary [45 C.F.R. §§ 164.502(b), 164.514(d)]. OCR HIPAA Privacy Fact Sheet, Revised 4 April 2003. Available online: https://www.hhs.gov/sites/default/files/ocr/privacy/hipaa/understanding/coveredentities/minimumnecessary.pdf (accessed on 3 September 2026).
- Weait, Matthew. 2019. Limit Cases: How and Why We Can and Should Decriminalise HIV Transmission, Exposure, and Non-Disclosure. Medical Law Review 27: 576–96. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yang, Dean, James Allen, IV, Arlete Mahumane, James Riddell, IV, and Hang Yu. 2023. Knowledge, Stigma, and HIV Testing: An Analysis of a Widespread HIV/AIDS Program. Journal of Development Economics 160: 102958. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yıldırım, Engin, and Serdar Gülener. 2018. Anayasa Mahkemesi Kararlarında Uluslararası ve Karşılaştırmalı Hukuka Yapılan Atıflar: Ampirik Bir Analiz [Citations to International and Comparative Law in the Decisions of the Constitutional Court: An Empirical Analysis]. Ankara Üniversitesi Hukuk Fakültesi Dergisi 67: 105–44. [Google Scholar]
- Yu, Hang. 2023. Social Stigma as a Barrier to HIV Testing: Evidence from a Randomized Experiment in Mozambique. Journal of Development Economics 161: 103035. [Google Scholar] [CrossRef] [Scilit]
| 1 | Türkiye Cumhuriyeti Anayasası (Constitution of the Republic of Türkiye), Law No. 2709, Resmî Gazete, 9 November 1982, No. 17863, as amended. Available online: https://www.mevzuat.gov.tr/mevzuatmetin/1.5.2709.pdf (accessed on 2 October 2026). |
| 2 | Sosyal Sigortalar ve Genel Sağlık Sigortası Kanunu (Social Insurance and General Health Insurance Act), Law No. 5510, adopted 31 May 2006, Resmî Gazete, 16 June 2006, No. 26200. |
| 3 | Debts were unlawfully raised against welfare recipients by automated annual income averaging, causing documented psychological harm and, in some cases, suicide: Royal Commission into the Robodebt Scheme, Report (Commonwealth of Australia 2023), vol. 2, chap. 10; see also Alston (2019, para. 23). |
| 4 | Türkiye Cumhuriyeti Anayasasının Bazı Maddelerinde Değişiklik Yapılması Hakkında Kanun, Law No. 5982, Resmî Gazete, 13 May 2010, No. 27580, which introduced Article 20(3) and the individual application mechanism. |
| 5 | Anayasa Mahkemesi (Constitutional Court), İlker Başer ve Diğerleri, B. No: 2013/1943, 9 September 2015. |
| 6 | Anayasa Mahkemesi (Constitutional Court), Kâmil Selman Bilgin ve Diğerleri, B. No: 2021/52365, 28 January 2026. Available online: https://kararlarbilgibankasi.anayasa.gov.tr/kbb/pages/search/BireyselBasvuru?id=a2JiOmU4N2M3MmIzLWVmZjUtNDc0My04NWNjLWRjYjJiYmQ1YzczYw&type=BireyselBasvuru (accessed on 25 July 2026). |
| 7 | Kişisel Verilerin Korunması Kanunu (Personal Data Protection Law), Law No. 6698, adopted 24 March 2016, Resmî Gazete, 7 April 2016, No. 29677. Available online: https://www.mevzuat.gov.tr/mevzuatmetin/1.5.6698.pdf (accessed on 2 October 2026). Article 15(5) provides that where an examination establishes a breach the Board decides that the unlawfulness be remedied by the controller and notifies the parties, and that the decision is to be complied with without delay and within thirty days of notification at the latest; Article 18 sets the administrative fines. |
| 8 | Umumi Hıfzıssıhha Kanunu (General Public Health Law), Law No. 1593, adopted 24 April 1930, Resmî Gazete, 6 May 1930, No. 1489. |
| 9 | Unlike legal systems with a single codified administrative procedure act—Germany’s Verwaltungsverfahrensgesetz is the paradigmatic example—Turkish administrative law has no general procedural code of that kind. |
| 10 | Kişisel Sağlık Verileri Hakkında Yönetmelik (Regulation on Personal Health Data), Resmî Gazete, 21 June 2019, No. 30808. |
| 11 | The Constitutional Court uses international instruments to specify this duty rather than to supply it: in review of constitutionality they rank above statutes but below the Constitution and operate as a supporting standard of review (‘destek ölçü norm’)—in the Court’s own formulation, one ‘not relied upon in the review of constitutionality but taken into account in the assessment’ (E. 1996/15, K. 1996/34, 23 September 1996)—the Convention system being the most frequently cited among them (Yıldırım and Gülener 2018, p. 119). The texts relied on are these. Constitution of the Republic of Türkiye, Art. 5: the fundamental aims and duties of the state include striving for the removal of the political, economic and social obstacles that restrict fundamental rights and freedoms in a manner incompatible with the principles of justice and of the social state governed by the rule of law, and providing ‘the conditions required for the development of the individual’s material and spiritual existence’ (‘insanın maddi ve manevi varlığının gelişmesi için gerekli şartları hazırlamaya çalışmaktır’). Art. 17(1): ‘Everyone has the right to life and the right to protect and improve his/her corporeal and spiritual existence.’ The Turkish text of both provisions reads ‘maddi ve manevi varlık’; the official English translation renders it ‘material and spiritual’ in Article 5 and ‘corporeal and spiritual’ in Article 17. 1961 Constitution, Art. 49: ‘Devlet, herkesin beden ve ruh sağlığı içinde yaşıyabilmesini ve tıbbî bakım görmesini sağlamakla ödevlidir’—the state is under a duty to ensure that everyone may live in physical and mental health and receive medical care—the first provision on health in Turkish constitutional history. Türkiye ratified the International Covenant on Civil and Political Rights and the International Covenant on Economic, Social and Cultural Rights on 23 September 2003; under Article 90, final paragraph, of the Constitution, added by Law No. 5170 of 7 May 2004, duly ratified agreements concerning fundamental rights and freedoms prevail where a statute conflicts with them. Human Rights Committee, General Comment No. 16 (1988) on Article 17 ICCPR, sct. 10, requires that the gathering and holding of personal information on computers and data banks, ‘whether by public authorities or private individuals or bodies, must be regulated by law’, and that effective measures ensure such information does not reach persons not authorized to receive, process and use it. Committee on Economic, Social and Cultural Rights, General Comment No. 14 (2000), UN Doc. E/C.12/2000/4, sct. 33, frames the right to health as imposing obligations to respect, protect and fulfil; sct. 12(b)(iv) provides that accessibility of health information ‘should not impair the right to have personal health data treated with confidentiality’. |
| 12 | Anayasa Mahkemesi (Constitutional Court), Bülent Kaya, B. No: 2013/2941, 11 May 2016. |
| 13 | Z v Finland, App no 22009/93, European Court of Human Rights, 25 February 1997. |
| 14 | I v Finland, App no 20511/03, European Court of Human Rights, 17 July 2008. The applicant, an HIV-positive nurse treated at the hospital where she worked, alleged that colleagues had consulted her records; because the system logged no viewing, it could not be established who had done so, and the Court held the deficient design itself to breach Article 8 (sct. 44, 47). |
| 15 | Anayasa Mahkemesi (Constitutional Court), T.A.A., B. No: 2014/19081, 1 February 2017. |
| 16 | The interest Article 17 protects in this setting is dignity in its informational aspect: what an unauthorized disclosure takes from the person is not secrecy as such but the standing to present themselves to others on their own terms. That characterization is ours rather than the Court’s, and it names the interest at stake without adding a separate ground of argument. |
| 17 | Ceza Muhakemesi Kanunu ile Bazı Kanunlarda Değişiklik Yapılmasına Dair Kanun, Law No. 7499, adopted 2 March 2024, Resmî Gazete, 12 March 2024, No. 32487, art. 33. |
| 18 | See Tietosuojavaltuutettu v Satakunnan Markkinapörssi Oy and Satamedia Oy, Case C-73/07, 16 December 2008, ECLI:EU:C:2008:727; Huber v Bundesrepublik Deutschland, Case C-524/06, 16 December 2008; Volker und Markus Schecke GbR and Eifert v Land Hessen, Joined Cases C-92/09 and C-93/09, 9 November 2010, ECLI:EU:C:2010:662; Digital Rights Ireland Ltd and Others, Joined Cases C-293/12 and C-594/12, 8 April 2014, ECLI:EU:C:2014:238; Tele2 Sverige AB and Watson and Others, Joined Cases C-203/15 and C-698/15, 21 December 2016. The Court’s usage is not uniform: (Rodin 2025, p. 184), writing in a scholarly rather than a judicial capacity, observes that necessity and proportionality are applied incoherently across the digital-rights case law. |
| 19 | Criminal Proceedings against V.S., Case C-205/21, 26 January 2023, ECLI:EU:C:2023:49, sct. 118. The case was decided under Directive (EU) 2016/680 (Law Enforcement Directive); for scholarly discussion of sensitive-data processing under that instrument, see Naarttijärvi (2023). |
| 20 | Regulation (EU) 2016/679 (General Data Protection Regulation), arts 25, 35, 36 and 9(2)(g). Available online: https://eur-lex.europa.eu/eli/reg/2016/679/oj (accessed on 4 September 2026). |
| 21 | Health Insurance Portability and Accountability Act of 1996 (HIPAA), Pub. L. 104–191; HIPAA Privacy Rule, 45 C.F.R. sct. 164.502(b), 164.514(d) (minimum necessary standard). Available online: https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-C/part-164/subpart-E/section-164.514 (accessed on 4 September 2026). |
| 22 | Regulation on Personal Health Data, Resmî Gazete, 21 June 2019, No. 30808, repealing the 2016 Regulation. |
| 23 | Danıştay 15. Dairesi (Council of State, 15th Chamber), E. 2016/10488 and E. 2016/10500, 6 July 2017 (staying arts 5(5) and 14 of the 2016 Regulation); Danıştay 15. Dairesi, E. 2018/251, 9 October 2018 (staying the amending instrument of 24 November 2017 in its entirety). The texts of both decisions, and of Danıştay 15. Dairesi, E. 2016/10500, 6 July 2017, are reproduced in (Kaya and Taştan 2025). The 2016 Regulation was the Kişisel Sağlık Verilerinin İşlenmesi ve Mahremiyetinin Korunması Yönetmeliği, Resmî Gazete, 20 October 2016, No. 29863. Article 6(4) of the 2019 Regulation, on e-Nabız security settings, was subsequently annulled by Danıştay 10. Dairesi, E. 2019/10544, K. 2023/5655, 17 October 2023, and redrafted by the amendment of 3 December 2025; the annulment is recorded in the consolidated text of the Regulation, immediately following Article 6(4). |
| 24 | Kişisel Sağlık Verileri Hakkında Yönetmelikte Değişiklik Yapılmasına Dair Yönetmelik, Resmî Gazete, 3 December 2025, No. 33096. Article 5 of the amending instrument replaced arts 6(2)–(5) of the 2019 Regulation; article 12 replaced the words ‘their authorization’s, if any, shall be revoked’ in art. 21(2) with ‘the disciplinary authority shall take the necessary measures’. |
| 25 | Kişisel Sağlık Verileri Hakkında Yönetmelikte Değişiklik Yapılmasına Dair Yönetmelik, Resmî Gazete, 4 July 2026, No. 33300, arts 1–3 (amending the heading of Chapter Four and Article 13(2), and inserting Article 13/A). Article 13/A(2) provides that, saving provisions of other legislation, diagnoses recorded in the past may not be taken as a basis in any process, recruitment processes above all. Available online: https://www.resmigazete.gov.tr/eskiler/2026/07/20260704-9.htm (accessed on 2 October 2026). |
| 26 | Kişisel Verileri Koruma Kurulu (Personal Data Protection Board), K. No: 2021/962, 21 September 2021. Available online: https://www.kvkk.gov.tr/Icerik/7074/2021-962 (accessed on 4 September 2026). |
| 27 | Kişisel Verileri Koruma Kurulu, K. No: 2023/695, 2 May 2023. Available online: https://www.kvkk.gov.tr/Icerik/7754/2023-695 (accessed on 4 September 2026). |
| 28 | Kişisel Verileri Koruma Kurulu, K. No: 2021/761, 6 August 2021. Available online: https://www.kvkk.gov.tr/Icerik/7137/2021-761 (accessed on 4 September 2026). |
| 29 | Kişisel Verileri Koruma Kurulu, K. No: 2020/355, 7 May 2020. The Board fined the pharmacy 60,000 TL and referred the pharmacist’s spouse to the public prosecutor under Article 136 of the Turkish Criminal Code (Law No. 5237). Available online: https://www.kvkk.gov.tr/Icerik/6767/2020-355 (accessed on 4 September 2026). |
| 30 | Kişisel Verileri Koruma Kurulu, K. No: 2023/1130, 6 July 2023. Available online: https://www.kvkk.gov.tr/Icerik/7770/2023-1130 (accessed on 4 September 2026). |
| 31 | Kişisel Verileri Koruma Kurulu, K. No: 2018/143, 5 December 2018. Available online: https://www.kvkk.gov.tr/Icerik/5364/2018-143 (accessed on 4 September 2026). |
| 32 | Kişisel Verileri Koruma Kurulu, K. No: 2023/1130, 6 July 2023, defining the information covered by pharmacist confidentiality as that whose disclosure would cause the patient to be ‘ayıplanacağı, dışlanacağı’—literally, to be shamed and excluded. |
| 33 | Kişisel Verileri Koruma Kurulu, K. No: 2022/594, 22 June 2022 (administrative fine of 75,000 TL). Available online: https://www.kvkk.gov.tr/Icerik/7566/2022-594 (accessed on 4 September 2026). |
| 34 | Kişisel Verileri Koruma Kurulu, K. No: 2023/767, 11 May 2023. Available online: https://www.kvkk.gov.tr/Icerik/7758/2023-767 (accessed on 4 September 2026). |
| 35 | On 3 September 2026 we searched the Board’s published decision summaries for the keyword ‘sağlık’; the search returned 67 results and we read all of them. They span 5 December 2018 to 26 December 2024. The endpoint is not a cut-off we imposed: the most recent summary the search returned was dated 26 December 2024, and the published record, as it stood at the date of the search, went no further. The set is heterogeneous, because Article 6(1) KVKK enumerates ‘health and sexual life’ among the special categories and any decision citing that provision is returned: it includes decisions on trade union membership, building-entry biometrics, credit registry records and library membership alongside those, such as the hospital and pharmacy decisions above, that concern health data substantively. The search was run on the Board’s decision-summary search facility at https://www.kvkk.gov.tr/search-decision (section identifier 5406) with the query string ‘sağlık’; the count of 67 is the figure the facility reported. The facility was returning no results when this revision was completed, so the corpus is reported as observed on the date of the search, and the log is supplied as Table S1. The individual summaries are unaffected: each remains retrievable at its own permanent address of the form https://www.kvkk.gov.tr/Icerik/[id]/[year]-[number], and the eight analyzed here are cited at those addresses above. |
| 36 | The point is developed in the digital health literature: once health data enters a centralized platform the individual loses practical control over it (Prentice 2013, quoted in Lupton 2014, p. 1354). |
| 37 | Article 15(7) of Law No. 6698: ‘Kurul, telafisi güç veya imkânsız zararların doğması ve açıkça hukuka aykırılık olması hâlinde, veri işlenmesinin veya verinin yurt dışına aktarılmasının durdurulmasına karar verebilir’—where damage that is difficult or impossible to repair arises and there is manifest unlawfulness, the Board may decide to halt the processing of the data or its transfer abroad. The power proposed here differs from it in trigger and in scope: it would respond to repeated failure to meet the technical requirements of Article 12 rather than to imminent damage, and would reach only the controller’s authorization to process very-high-risk data. |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.