Secure Digital Rights Management in Gamified Personal Health Promotion Applications Using Attribute-Based Encryption

Abstract

The rising prevalence of diseases linked to factors such as excessive alcohol and tobacco use, sedentary lifestyles, and poor nutrition has led to a greater focus on Personal Health Promotion (PHP) as a preventive measure. PHP emphasizes improving quality of life and well-being, driven by advances in medical technology, including virtual and augmented reality. However, as PHP systems grow in popularity, concerns over personal data security, such as account theft and information breaches, have become increasingly important. To address these concerns, this study proposes a game-based personalized health promotion system that integrates Key-Policy Attribute-Based Encryption (KP-ABE), key insulation, time-bound encryption. These methods ensure data security through hierarchical access control and dynamic time management, protecting personal health records and exercise prescriptions. The system also incorporates key insulation to enhance the security of private keys. This multi-layered security approach provides a robust solution for safeguarding sensitive data within PHP systems while accommodating dynamic subscription needs and legal access requirements.

1. Introduction

As the global population ages and chronic diseases become more common, traditional healthcare systems face increasing pressure. Smart healthcare systems offer a solution by using technologies like the Internet of Things (IoT), Artificial Intelligence (AI), big data, and blockchain to improve medical service efficiency and address system challenges [1]. These systems provide real-time health monitoring and data analysis, helping individuals manage their health and assisting healthcare professionals in making accurate diagnoses and treatment decisions. For the elderly and people with chronic conditions, smart healthcare offers continuous support, reducing hospital visits and lowering healthcare costs by preventing unnecessary hospitalizations and emergencies [2]. Wearable devices and smartphones play a key role by enabling real-time health tracking, increasing awareness, and supporting personalized health promotion through remote monitoring [3].

Personalized services are essential for the effectiveness of smart healthcare systems. Since individuals have unique health needs, personalized models create tailored health promotion plans based on personal health data, lifestyle, and genetic background. Personalized medicine focuses not only on treatment but also on disease prevention and health promotion [4]. For example, smart healthcare systems can provide fitness recommendations, dietary advice, and medication reminders based on real-time health data, making health management more accurate and convenient [5]. This approach empowers users to take control of their health, leading to better outcomes.

Gamification is a powerful tool in smart healthcare, using game-like elements to engage users and encourage health management. It motivates users to change behaviors and improve health outcomes [6]. In personalized health promotion, gamification adapts activities to an individual’s health status, providing real-time feedback, challenges, and rewards to sustain interest. For instance, platforms may assign daily health tasks based on user data and offer rewards for participation. A diabetes management app might encourage blood glucose monitoring through challenges, rewarding users for consistent engagement [5].

There are many examples of gamification in smart healthcare. For the elderly, health promotion systems combine cognitive challenges with physical activities to maintain mental and physical health. These systems use wearable devices or mobile apps to monitor activities and adjust difficulty based on real-time health data [1]. For individuals with obesity, health management apps offer gamified exercise challenges and diet tracking. These apps reward users with virtual prizes for reaching daily goals. Research shows that gamified elements significantly improve user engagement and promote behavior change [2].

In the realm of health-promoting games, the Steam platform offers a wide range of virtual reality (VR) titles that seamlessly combine entertainment with fitness. Beat Saber, for instance, is a highly popular VR rhythm game in which players swing controllers to slice through incoming blocks to the beat of the music. This game provides an aerobic workout that elevates heart rate and improves coordination, allowing players to stay physically active while enjoying a gaming experience [7]. Similarly, the Meta Quest platform (formerly Oculus) focuses on VR fitness games like Supernatural and FitXR, which incorporate activities such as aerobic exercises, yoga, and boxing. These applications offer players a diverse array of fitness options in an immersive VR environment, making workouts more engaging and enjoyable [8]. Both platforms not only expand the available fitness options but also provide flexible solutions for both family and individual use, meeting various health needs and lifestyle preferences. In the context of VR health-promoting games, Digital Rights Management (DRM) plays a crucial role in managing and protecting digital content. As VR platforms increasingly offer personalized fitness experiences, DRM ensures that only authorized users can access and interact with these tailored services. This security measure is particularly important for safeguarding subscription-based content, as it prevents unauthorized access, copying, and sharing of proprietary resources. By controlling content usage, DRM helps providers to maintain operational efficiency and secure distribution, while also enabling users to enjoy their personalized experiences within the VR environment without compromising content integrity.

While health promotion systems provide personalized services, managing tailored content for each user is challenging and may reduce efficiency. A Digital Rights Management (DRM) approach addresses this by securely delivering personalized services based on user subscriptions or access rights. DRM ensures that only authorized users can access digital content and controls actions such as copying, printing, sharing, and editing. It enforces usage licenses and uses encryption to ensure that only authorized devices or applications can decrypt the content [9,10]. Secure Digital Rights Management (SDRM) enhances this security and is widely used in enterprises and government agencies. SDRM uses advanced encryption, access control, and user tracking to protect digital resources in distributed environments [6].

In healthcare, information security is critical for electronic prescriptions and personalized health services. Managing sensitive health data is complex due to the diverse roles of individuals accessing and handling it. Robust protection and oversight are essential. Key-Policy Attribute-Based Encryption (KP-ABE) offers a solution by enabling fine-grained, hierarchical access control using specific attributes for authorization. However, managing dynamic access permissions over time is a significant challenge. Adding time-bound access control to KP-ABE strengthens security by limiting data access to specific time frames. This prevents unauthorized access from expired or outdated permissions. This approach improves data protection and provides flexibility to meet the changing security needs of healthcare environments.

The research problems of this study are described as follows:

(1) Health promotion systems should focus on not only health promotion functions but also on digital rights management and the protection of sensitive personal records. This study should consider these issues and design concrete cryptographic solution to protect data security and intelligent property of health promotion system.

(2) Elaborating on cryptographic solutions to access control or oversee privilege management for health promotion systems, this study will design a Key-Policy Attribute-Based Encryption (KP-ABE) with key insulation. Previously proposed KP-ABE schemes do not provide key insulation. A KP-ABE scheme without key insulation might not achieve security requirements if the key is compromised or the secret key is not updated for some period. Hence, a KP-ABE scheme with key insulation can achieve higher security promise for secret key protection. That is, such a scheme can force users or the system to update the secret key periodically, but the corresponding public key or public information can remain the same. This security feature is useful and important for health promotion systems.

(3) A health promotion system should not only consider privilege management but also provide security protection for access control for a legal period or the time-limitation of access control. Hence, we should provide a time-bound service for health promotion systems so that valid access control rights are valid in a legal time period. Different from key insulation, a time-bound service focuses on limiting valid access privilege. This study will provide a secure time-bound cryptographic protection mechanism to limit the access control privilege for legal users or authorized services.

(4) The proposed cryptographic solutions should be based on elliptic curve cryptography (ECC) to achieve higher security protection and better performance.

With the growth of the gaming industry, managing game copyrights and controlling access rights has become critical. Traditional encryption methods provide basic encryption and decryption, granting full access after decryption. This approach lacks granular control over game permissions, leading to digital rights losses. Additionally, leaked game keys allow attackers to modify game content, such as altering parameters in health promotion games, potentially endangering users. Our method addresses this issue by offering on-demand dynamic subscriptions with DRM protection.

Personalized health records (PHR) are now an essential part of daily life, but they face significant information security risks. If PHR data are stolen, it can lead to privacy breaches, identity theft, or data resale. Traditional PHR protection methods focus on defending against external attacks but fail to prevent internal data theft. Moreover, standard encryption cannot enforce effective access control and often neglects key security, leading to irreversible damage if private keys are leaked.

This study combines encrypted games and PHR for health promotion. During encryption, each game targets specific health objectives, such as osteoporosis prevention or stroke risk reduction. Games are tailored to unique goals, allowing sites to request specific games based on their needs. Sites may select multiple games or combine them to suit their requirements.

The system employs a time-bound model for operation. It offers 15 games for subscription, addressing goals such as stroke risk reduction, general health promotion, and osteoporosis prevention. This model controls subscription permissions rather than gameplay duration, ensuring DRM protection. During encryption, the system validates the subscription period and encrypts games accordingly. Each game self-verifies during launch to confirm it remains within the subscription period.

PHR data, including game scores and user information, are generated during gameplay. Access to these data is managed using Key-Policy Attribute-Based Encryption (KP-ABE), allowing fine-grained access control for Site Case Managers (SCM) based on the games played. Monthly updates to the game and secret keys prevent data breaches from leaked keys.

For example, Site A may subscribe to games 1–5 for stroke risk reduction. If it later adds general health promotion, it would subscribe to games 1–10. An SCM with access to game 1 (G1) can only view data from G1, while access attempts for other games will fail due to mismatched KP-ABE keys. A Physical Examination Center SCM, with broader access permissions, can view all game results to evaluate users and recommend exercise prescriptions.

This study proposes a personalized health promotion system that includes multiple security protection mechanisms. It uses cryptographic methods to enhance data protection. In this paper, we make the following contributions:

• Utilizes attribute-based encryption (ABE) to implement fine-grained, hierarchical access control, ensuring that only authorized users can access specific data based on their attributes;

• Incorporates a ciphertext update mechanism to dynamically manage changes in user permissions;

• Implements a key insulation mechanism to prevent data theft;

• Integrates a time-bound mechanism to regulate the duration of authorized access for users from different sectors or fields;

• Employs elliptic curve cryptography (ECC) to ensure data confidentiality;

• Verifies the correctness of the algorithm through mathematical formulas;

• Aims to protect personal information, health data, and exercise prescriptions while enabling dynamic permission management and preventing unauthorized access to system data.

2. Related Works

2.1. Personal Health Promotion

Personal Health Promotion (PHP) focuses on empowering individuals to manage their own health by adopting healthier behaviors and engaging in preventive health practices. PHP encompasses a range of activities designed to improve personal health, such as regular exercise, balanced nutrition, and stress management, which collectively help prevent chronic diseases and promote overall well-being. In practice, PHP strategies often utilize digital tools and resources, including mobile applications and cloud-based systems, which enable individuals to track their health data, receive personalized recommendations, and engage in self-directed care [11].

PHP and Personal Health Records (PHR) work synergistically to support personal health management by providing individuals with tools to record, monitor, and understand their health status. While PHP emphasizes promoting healthy behaviors, PHR serves as a platform for storing and managing personal health information, enabling individuals to make informed decisions based on comprehensive health data. Recent studies suggest that integrating PHP initiatives with PHR systems can help users track their health behaviors more effectively and provide healthcare professionals with valuable insights into patients’ health patterns, thus facilitating more personalized and preventive care [12].

In specific healthcare contexts, like managing chronic diseases, PHRs enable individuals to monitor health indicators consistently, align with PHP practices, and receive tailored support and guidance based on their unique health needs. This combination fosters a proactive approach to healthcare, enhancing patient engagement and improving health outcomes by empowering individuals to maintain regular health monitoring and self-management [13].

2.2. Key-Policy Attribute-Based Encryption

Key-Policy Attribute-Based Encryption (KP-ABE) is an extension of Attribute-Based Encryption (ABE), developed by Sahai and Waters. It addresses the limitations of traditional encryption in data access control by providing fine-grained management. Unlike ABE, which controls access using attributes, KP-ABE encrypts data based on attributes and embeds decryption policies into the keys. Users can decrypt data only if their attributes match the defined policies. This method was created to enhance data security and offer flexible access control, especially in distributed environments like cloud storage, where traditional Identity-Based Encryption (IBE) is insufficient.KP-ABE has advanced to support complex access policies, providing stronger security and efficient encryption with constant ciphertext sizes [14]. Its importance includes:

(1) Enabling fine-grained access control based on user attributes [15];

(2) Enhancing data privacy in untrusted cloud environments [16];

(3) Supporting customizable access policies for high-security applications, such as medical data sharing and cloud storage [14].

As shown in Figure 1, a policy can be created where access to encrypted medical records is restricted to individuals who are part of “Hospital A” and must hold both the roles of “Doctor” and “Nurse” simultaneously. Only users possessing all these attributes can decrypt the data, providing a clear application of KP-ABE in healthcare to ensure that sensitive data are only accessible to the appropriate personnel.

With the widespread adoption of internet and cloud computing technologies, secure data storage and sharing have become critical. Key-Policy Attribute-Based Encryption (KP-ABE) addresses the need for flexible access control and robust data security in several key scenarios:

(1) In cloud computing, KP-ABE allows data owners to encrypt their data and set decryption permissions through access policies, ensuring that data remains secure even on untrusted servers [17];

(2) As data privacy concerns grow, KP-ABE prevents unauthorized access by defining permissions based on user attributes;

(3) Traditional access control methods, such as Role-Based Access Control (RBAC), struggle with the complexity of distributed systems, whereas KP-ABE offers more efficient and flexible access control across diverse environments.

Key applications of KP-ABE include:

(1) Cloud storage, where data are encrypted before storage and only accessible to users with matching attributes [18];

(2) Sensitive data environments, such as healthcare and finance, where KP-ABE ensures that only authorized users can access critical information [14];

(3) IoT and distributed systems, where KP-ABE manages access rights based on attribute matching, ensuring only authorized devices or users can access data [19].

Key-Policy Attribute-Based Encryption (KP-ABE) has been widely studied for its security, applications, and technical challenges. Its security relies on number-theoretic assumptions, such as the bilinear Diffie–Hellman assumption. Researchers have worked to improve KP-ABE’s resilience against attacks, including using the Multi-Authority Model. This model distributes attribute management among multiple authorities to reduce risks from single points of failure and concentration of power [20]. KP-ABE is particularly effective in cloud computing, providing fine-grained access control that enhances data security and privacy in multi-user environments. It can also integrate with blockchain technology to secure and ensure the transparency of cloud-stored data. However, KP-ABE faces challenges, such as increased ciphertext size with more attributes, which reduces system efficiency. Recent advancements have addressed this by developing schemes with constant ciphertext sizes. These improvements lower storage and transmission costs while enhancing performance [14]. Future research aims to improve the efficient revocation of user permissions and manage dynamic attribute changes in real-time environments.

This study introduces novel security mechanisms specifically tailored for Personal Health Promotion (PHP) systems, advancing the Key-Policy Attribute-Based Encryption (KP-ABE) framework to enhance data protection while adapting to the dynamic needs of health promotion. While KP-ABE traditionally provides access control based on user attributes, our approach extends this capability by incorporating key insulation and time-bound access control, two essential features for PHP applications. Key insulation mitigates potential security risks by segmenting private key access into defined periods, thus preserving data integrity even if a key is compromised [12]. In parallel, time-bound access control automatically revokes permissions after predetermined intervals, ensuring that access aligns with the temporal requirements of health interventions [11]. These enhancements provide granular, flexible control over health data, significantly advancing beyond the traditional applications of KP-ABE discussed in prior studies [12]. By refining KP-ABE specifically for PHP systems, this approach establishes a more resilient and context-aware framework, effectively bridging the gap between preventive health strategies and secure data management within digital health environments.

Review of He et al.’s KP-ABE Scheme [21]

He et al. designed a key escrow-free KP-ABE scheme [21]. They address key escrow issues in IoT environments, while we introduce additional technical enhancements such as key insulation and time-bound encryption, which are especially important for sensitive data in networked environments. In this section, we present a brief review of Hong et al.’s scheme. Their approach is mainly divided into six steps. Central Authority Setup, User Registration, Proxy Authority Setup, Key Generation, and Encryption and Decryption.

The Central Authority Setup phase yields the system public parameters $\mathrm{PP}=\left(g,h,Z\right)$ and CA’s primary secret key ${\mathrm{MSK}}_{\mathrm{CA}}=\left(d_1,d_2,z\right).$ In the User Registration phase, the algorithm is performed by CA when a user submits his identity UID for registration. Then, the algorithm generates the user’s global secret key ${\mathrm{GSK}}_{\mathrm{uid}}=\left(g^x,g^y\right)$. In the Proxy Authority Setup phase, the algorithm is performed by PA. It computes PA’s public parameters ${\mathrm{PK}}_{\mathrm{PA}}=\left(f,Q,{\left\{U_j\right\}}_{j\in \mathcal{U}}\right)$ and the primary secret key ${\mathrm{MSK}}_{\mathrm{PA}}=\left(\alpha ,\beta ,\tau ,\theta ,{\left\{u_j\right\}}_{j\in \mathcal{U}}\right).$ In the Key Generation phase, the algorithm is also performed by PA. A user applies for a decryption key. The algorithm obtains the user’s UID and $g^{d_{2}y}$. Then, the algorithm yields the partial decryption key $\begin{array}{cc}\mathrm{SK}=& (K^{\prime }=g^{{\displaystyle \frac{\pi }{\tau +\sigma }}},{\left\{K_i=g^{{\displaystyle \frac{{\lambda }_i}{u_{\rho \left(i\right)}}}}\right\}}_{i=1}^{\mathcal{l}})K=g^{{\displaystyle \frac{\alpha }{\tau +\sigma }}}\cdot {\left(g^{d_{2}y}\right)}^{{\displaystyle \frac{1}{\tau +\sigma }}}\cdot {\left(g^{\theta -\sigma }\right)}^{{\displaystyle \frac{\pi }{\tau +\sigma }}},K_0=\sigma \\ & K^{\prime }=g^{{\displaystyle \frac{\pi }{\tau +\sigma }}},{\left\{K_i=g^{{\displaystyle \frac{{\lambda }_i}{u_{\rho \left(i\right)}}}}\right\}}_{i=1}^{\mathcal{l}})\end{array}$ and issues it to the corresponding user. After receiving $\mathrm{SK}$, the user combines it with his GSK to yield the complete decryption key $\begin{array}{cc}\mathrm{SK}=& (K^{\prime }=g^{{\displaystyle \frac{\pi }{\tau +\sigma }}},{\left\{K_i=g^{{\displaystyle \frac{{\lambda }_i}{u_{\rho \left(i\right)}}}}\right\}}_{i=1}^{\mathcal{l}})K=g^{{\displaystyle \frac{\alpha }{\tau +\sigma }}}\cdot {\left(g^{d_{2}y}\right)}^{{\displaystyle \frac{1}{\tau +\sigma }}}\cdot {\left(g^{\theta -\sigma }\right)}^{{\displaystyle \frac{\pi }{\tau +\sigma }}}\\ & D_0=g^x,D_0^{\prime }=g^y\\ & K_0=\sigma ,K^{\prime }=g^{{\displaystyle \frac{\pi }{\tau +\sigma }}},{\left\{K_i=g^{{\displaystyle \frac{{\lambda }_i}{u_{\rho \left(i\right)}}}}\right\}}_{i=1}^{\mathcal{l}}).\end{array}$ In the Encryption phase, the algorithm is performed by IoT nodes; ${\mathrm{PK}}_{\mathrm{PA}}$ is the public parameters of that area’s PA. The algorithm randomly selects $s\stackrel{R}{\leftarrow }{\mathbb{Z}}_p$ and generates the cyphertext $\begin{array}{cc}\mathrm{CT}=& ({\left\{C_j=U_j^s\right\}}_{j\in \gamma })\gamma ,\tilde{C}=M\cdot Z^s\cdot Q^s,C^{\prime }=g^s,C^{″}=f^s,C^{‴}=h^s\\ & {\left\{C_j=U_j^s\right\}}_{j\in \gamma }).\end{array}$. In the Decryption phase, the algorithm is performed by users. The algorithm calculates

$H={{\displaystyle \prod }}_{i\in I}e{\left(C_{\rho \left(i\right)},K_i\right)}^{{\omega }_i}={{\displaystyle \prod }}_{i\in I}e\left(U_{\rho \left(i\right)}^s,g^{{\displaystyle \frac{{\lambda }_i}{u_{\rho \left(i\right)}}}}\right)=e{\left(g,g\right)}^{s{{\displaystyle \sum }}_{i\in I}{\lambda }_i{\omega }_i}=e{{\left(g,g\right)}^{\left(\beta -\theta \pi \right)s}}^{{\omega }_i}$

$J=e\left({C^{\prime }}^{K_0}\cdot C^{″},K\cdot {K^{\prime }}^{K_0}\right)=e\left(g^{\left(\tau +\sigma \right)s},g^{{\displaystyle \frac{\alpha }{\tau +\sigma }}}{g}^{{\displaystyle \frac{d_{2}y}{\tau +\sigma }}}{g}^{{\displaystyle \frac{\left(\theta -\sigma \right)\pi }{\tau +\sigma }}}{g}^{{\displaystyle \frac{\sigma \pi }{\tau +\sigma }}}\right)=e{\left(g,g\right)}^{\left(\alpha +d_{2}y+\theta \pi \right)s}$

$L=e\left(C^{\prime },D_0\right)e\left(C^{‴},D_0^{\prime }\right)=e\left(g^s,g^x\right)e\left(h^s,g^y\right)=e{\left(g,g\right)}^{\left(x+d_{1}y\right)s}$

Finally, it retrieves the message $M={\displaystyle \frac{\tilde{C}}{L\cdot J\cdot H}}$.

Although they proposed a key escrow-free method, similar approaches cannot recover lost keys through a third party in the event of key loss. Additionally, this increases the burden of key management on users. Moreover, lost keys may expose users to potential attacks. To address these issues, we propose a key insulation method to mitigate these risks and enhance security.

2.3. Review of Bethencourt’s CP-ABE Scheme [22]

CP-ABE (Ciphertext-Policy Attribute-Based Encryption) is an encryption technique that allows message decryption based on the recipient’s attributes, such as position, department, or role [22]. It is commonly used to protect data privacy in distributed systems. For instance, a company can use CP-ABE to encrypt employee salary information so that only employees with specific roles can access it. By binding salary data to an access policy, decryption is restricted to those with matching attributes. CP-ABE can also safeguard other sensitive information, such as medical records or financial transactions.

The Ciphertext-Policy Attribute-Based Encryption process typically consists of four stages:

1. Initialization Phase: system parameters and public keys are generated.

2. Key Generation Phase: secret keys are generated for users based on their attributes.

3. Encryption Phase: data are encrypted with an access policy.

4. Decryption Phase: users with matching attributes decrypt the ciphertext to access the data.

In the initialization phase, choose two secret parameters $\alpha ,\beta \in Z_p$, create public key PK $=$ $\left(G_0,g,h=g^{\beta },f=g^{-\beta },e{\left(g,g\right)}^{\alpha }\right)$, and a master secret key MK $=\left(\beta ,g^{\alpha }\right)$.

This algorithm takes a set $S$ as input and outputs a corresponding set of keys (public key and private key) for $S$. First, a random value $r\in Z_p$ is selected. Then, for each attribute $j\in S$, a random $r_j\in Z_p$ is selected. The private key $SK$ is then generated and output as follows: $SK=\left(D=g^{{\displaystyle \frac{\alpha +r}{\beta }}}\forall j\in S:D_j=g^{r}H{\left(j\right)}^{r_j},D_j^{\prime }=g^{r_j}\right)$.

In the Encryption Phase, this algorithm encrypts a message $M$ based on an access policy $\mathcal{U}$. The process begins by converting the access policy into an access tree. For each node $R$ in the tree, a polynomial $p_R$ is defined, with degree $t_R-1$. First, a secret random value $s\in Z_p$ is selected, and $p_R\left(0\right)=s$ is set. Then, $t_R-1$ random constants are chosen to complete the polynomial $p_R$. For any other node $x$, $p_x\left(0\right)=p_{parent\left(x\right)}\left(index\left(x\right)\right)$ is set, where $parent\left(x\right)$ is the parent node of $x$ and $index\left(x\right)$ is the index of $x$ under its parent. Similarly, $t_x-1$ random constants are chosen to complete the polynomial $p_x$. Finally, let $S$ be the set of leaf nodes in the access tree. The ciphertext is generated as: CT = $(\mathcal{U},\overline{C}=Me{\left(g,g\right)}^{\alpha s},C=h^s,\forall x\in X:C_x=g^{p_x\left(0\right)},C_x^{\prime }=H{\left(att\left(x\right)\right)}^{p_x\left(0\right)}$).

The ciphertext CT can be decrypted using the recipient’s private key SK. If the recipient’s attribute set S satisfies the access structure, the ciphertext CT can be decrypted to retrieve the message M. The detailed recursive decryption algorithm is as follows:

For each non-leaf node $x$, let $i=attr\left(x\right)$.

If $i\in S$: $\mathrm{DecryptNode}\left(CT,SK,x\right)=e\left(D_i,C_x\right)·e{\left(D_i^{\prime },C_x^{\prime }\right)}^{-1}=e\left(g^{r}H{\left(i\right)}^{r_i},h^{p_x\left(0\right)}\right)·e{\left(g^{r_i},H{\left(i\right)}^{p_x\left(0\right)}\right)}^{-1}=e{\left(g,g\right)}^{r{p}_x\left(0\right)}$.

If $i\notin S$: DecryptNode(CT,SK,$x)=$ ⊥ otherwise

If $x$ is a non-leaf node, the recursive case is considered as follows: let $z$ be a child node of $x$. Call DecryptNode(CT,SK,$x$), which outputs $F_Z$. Define $S_x$ as an arbitrary set of size $t_x$. Collect child nodes $z$ such that $F_Z\ne \perp$. If no such $S_x$ exists, then DecryptNode(CT,SK,$x)=$ ⊥ otherwise; otherwise, proceed with the decryption which $p_x\left(0\right)=\begin{array}{c}s={{\displaystyle \sum }}_{i\in I}{\lambda }_{i}p\left(i\right)\end{array}\mathrm{and}\begin{array}{c}{\lambda }_i={\displaystyle \frac{{{\displaystyle \prod }}_{j\in I\setminus \left\{i\right\}}j}{{{\displaystyle \prod }}_{j\in I\setminus \left\{i\right\}}\left(j-i\right)}}\end{array}$, $I=\left\{1,\dots ,n\right\}$:

$$\begin{gathered} F_z={\displaystyle \prod }_{z\in S_x}{F_z}^{{\Delta }_{i,{S^{\prime }}_x}\left(0\right)},wherei=index\left(z\right),{S^{\prime }}_x=\left\{index\left(z\right):z\in S_x\right\}={\displaystyle \prod }_{z\in S_x}{\left(e{\left(g,g\right)}^{r{p}_z\left(0\right)}\right)}^{{\Delta }_{i,{S^{\prime }}_x}\left(0\right)} \\ ={\displaystyle \prod }_{z\in S_x}{\left(e{\left(g,g\right)}^{p_{parrent\left(z\right)}\left(index\left(z\right)\right)}\right)}^{{\Delta }_{i,{S^{\prime }}_x}\left(0\right)}={\displaystyle \prod }_{z\in S_x}{\left(e{\left(g,g\right)}^{r{p}_x\left(i\right)}\right)}^{{\Delta }_{i,{S^{\prime }}_x}\left(0\right)}=e{\left(g,g\right)}^{r{p}_x\left(0\right)} \end{gathered}$$

Finally, the decryption process works its way back from the leaf nodes to the root node. R = $e{\left(g,g\right)}^{r{p}_R\left(0\right)}$ and calculate $DecrypNode\left(CT,SK,x\right)=e{\left(g,g\right)}^{r{p}_R\left(0\right)}=e{\left(g,g\right)}^{rs}$. Obtain plain text:

$M={\displaystyle \frac{\tilde{C}}{e\left(C,D\right)/e{\left(g,g\right)}^{rs}}}={\displaystyle \frac{\tilde{C}}{e\left(h^s,g^{\frac{\alpha +r}{\beta }}\right)/e{\left(g,g\right)}^{rs}}}={\displaystyle \frac{\tilde{C}}{e\left(g^{\beta s},g^{\frac{\alpha +r}{\beta }}\right)/e{\left(g,g\right)}^{rs}}}={\displaystyle \frac{\tilde{C}}{e{\left(g,g\right)}^{s\left(\alpha +r\right)}/e{\left(g,g\right)}^{rs}}}={\displaystyle \frac{\tilde{C}}{e{\left(g,g\right)}^{\alpha s}}}$

2.4. KP-ABE and CP-ABE Summary

As shown in Table 1, the two algorithms exhibit a complementary relationship. The comparison between KP-ABE and CP-ABE highlights their key differences in terms of attribute matching, access policy, control, computational requirements, and security. In KP-ABE, the attributes associated with the ciphertext are matched against the access policy embedded in the decryption key, limiting the data owner’s control over who can access the data. Conversely, CP-ABE allows the data owner to embed the access policy directly into the ciphertext, enabling them to decide who obtains access based on attributes.

While KP-ABE requires lower computational resources, it provides weaker resistance to collusion attacks, as multiple users may combine their keys to bypass access restrictions. On the other hand, CP-ABE, though more computationally intensive, offers stronger security against such attacks, making it more suitable for scenarios requiring fine-grained and secure access control.

2.5. Key-Insulation

Key-insulation is a crucial encryption technique designed to mitigate the risks associated with key exposure, particularly in environments where devices, such as mobile phones or cloud servers, are vulnerable to attacks. Its importance lies in ensuring data protection and key management, even on devices with limited security. Key-insulation works by dividing the private key into distinct time periods, with each period having its own unique private key that must be regularly updated. Only one private key is active and usable within a given period, and it is applicable solely for that specific timeframe. Therefore, if a private key is compromised, the impact is limited to that period, ensuring that the security of other periods remains unaffected. Periodic updates to vulnerable sub-keys using a secure master key improve long-term security [23]. The advantages of key-insulation include: (1) resistance to key exposure, ensuring system security even if part of the key is compromised; (2) protection against key leakage, keeping data secure during multiple attack phases; (3) suitability for mobile and IoT devices, allowing secure encryption in untrusted environments.

The need for key-insulation technology arises from the challenge of mitigating key exposure risks as encryption becomes more widely used in systems and devices. Traditional key management methods often fail to ensure long-term security in environments such as IoT and smart devices. Key-insulation solves this issue by enabling periodic key updates, which improve security over time. This approach is particularly effective for mobile device encryption, where the risk of key exposure is higher due to insecure environments. Regular key updates reduce the likelihood of attackers compromising private data. In cloud computing, where data are stored on third-party servers, key-insulation allows secure key updates without disrupting operations, ensuring data security in high-risk environments.

Key-insulation techniques improve encryption security by separating the master key from sub-keys. This ensures that if a sub-key is compromised, the master key stays secure. Several methods have been developed to improve key-insulation effectiveness. (1) Key segmentation and update methods divide the key into segments and refresh them regularly. This helps reduce the risk of key exposure. Regular updates to sub-keys using the master key help mitigate potential compromises. (2) Hybrid key-insulation models combine key-insulation with other cryptographic techniques, such as Attribute-Based Encryption (ABE), to enhance security and flexibility. Research shows that integrating key-insulation with modern cryptographic methods improves both performance and security. In high-security environments, combining key-insulation with multi-party computation strengthens protection against key exposure [24].

2.6. Review of Hsu et al.’s Key-Insulated Scheme [25]

Hsu et al. designed an Identity-based Key-insulated Convertible Multi-authenticated Encryption Scheme [25]. This section primarily discusses the key insulation aspect, providing a brief overview of key insulation registration and key update processes. There are three characters, User, PKG, and Helper. While PKG generates the secret key for the user, and Helper generates the help key for the user. The User’s public key is ${\sigma }_{A_j}=e\left(P_{TA},H_1(I{D}_{A_j})\right)$ and secret key represented as $S_{A_j,0}$ = $s{H}_1$($I{D}_{A_j}$) + $w{H}_1\left(I{D}_{A_j},0\right)$. PKG’s public key is P_TA = sP, P_HK = wP, and secret key is s, w. The Helper can be PKG or another unit; their main mission is to generate the help key. Therefore, in some scenarios, it may be more suitable for the Helper not to be the PKG, in order to reduce the computational load on the PKG.

In the key insulation registration phase, they let $A=\left\{A1,A2,\dots ,An\right\}$ as a group of Users, in which n represents the number of Users. The User will send $I{D}_{A_j}$ to PKG, and PKG will generate the User’s secret key $S_{A_j,0}$ = $s{H}_1$($I{D}_{A_j}$) + $w{H}_1\left(I{D}_{A_j},0\right)$. Then, the User will generate their public key ${\sigma }_{A_j}=e\left(P_{TA},H_1(I{D}_{A_j})\right)$. We can see the detailed transmission process and steps in Algorithm RW1, which represent as the first algorithm in related work.

Algorithm RW1 Key_Reg
User $A_j$	PKG
1. Group A = {A1, A2, …, An} the number of users n.
2. Let the id of $User{A}_j$as $I{D}_{A_j}$
3. $I{D}_{A_j}$
	4. H1($I{D}_{A_j}$)
	5. $S_{A_j,0}$ = $s{H}_1$($I{D}_{A_j}$) + $w{H}_1\left(I{D}_{A_j},0\right)$
	6. $S_{A_j,0}$
7. ${\sigma }_{A_j}=e\left(P_{TA},H_1(I{D}_{A_j})\right)$

In the key update phase, the Helper generates a help key $H{K}_{A_j,i}=w\left[H_1\left(I{D}_{A_j},i\right)-H_1\left(I{D}_{A_j},i-1\right)\right]$ and send it to the User. The User will update secret key by the help key $S_{A_j,i}=S_{A_j,i-1}+H{K}_{A_j,i}$ = $s{H}_1$($I{D}_{A_j}$) + $w{H}_1\left(I{D}_{A_j},i-1\right)+w\left[H_1\left(I{D}_{A_j},i\right)-H_1\left(I{D}_{A_j},i-1\right)\right]$. After Key_Upd algorithm, the User’s secret key will update from $S_{A_j,0}$ to $S_{A_j,i}$. We can see the detailed transmission process and steps in Algorithm RW2.

Algorithm RW2 Key_Upd
Helper	User $A_j$
1. $H{K}_{A_j,i}=w\left[H_1\left(I{D}_{A_j},i\right)-H_1\left(I{D}_{A_j},i-1\right)\right]$(time period i ∈{1, …, N})
2. $H{K}_{A_j,i}$
	3. $S_{A_j,i}=S_{A_j,i-1}+H{K}_{A_j,i}$= $s{H}_1$($I{D}_{A_j}$) + $w{H}_1\left(I{D}_{A_j},i-1\right)+$$w\left[H_1\left(I{D}_{A_j},i\right)-H_1\left(I{D}_{A_j},i-1\right)\right]$

2.7. Review of Hong et al.’s Key-Insulated CP-ABE [26]

Hong et al. construct a key-insulated ciphertext policy attribute-based encryption [26]. Their algorithm includes five parts, Setup, Initial private key generation, Key update, and Encrypt and Decrypt.

In the Setup phase, the algorithm takes φ as input and the output system public parameter PK, master key MK, and helper key HK.

In the Initial private key generation phase, the algorithm inputs the system public key PK, master key $MK$, user’s id with attribute set $\left\{A_i\right\}$, and initial time period $T{P}_0$. The attribute authority (AA) outputs the user’s secret key $S{K}_{id,T{P}_0}$.

In the Encrypt phase, it inputs system public key PK, plaintext M, and encryption policy $\mathsf{\gamma }$, and outputs ciphertext CT.

In the Decrypt phase, it takes CT and the receiver’s temporal private key as input. It outputs the plaintext M.

In the Key update phase, there are two time periods $T{P}_n$ and $T{P}_{n+1}$, an attribute set $\left\{A_i\right\}$, and $HK$. The algorithm generates $U{P}_{i,T{P}_{n+1}}$. Users update their secret key by $U{P}_{i,T{P}_{n+1}}$. The final key obtained by the user will follow as $S{K}_{id,T{P}_{n+1}}=S{K}_{id,T{P}_n}·U{P}_{i,T{P}_{n+1}}$.

KI-CPABE-KEA focuses on addressing key leakage with features like accountability, as well as forward and backward security. However, it lacks fine-grained control over data access within specific time periods. Our scheme builds on its key insulation foundation by incorporating a time-restriction mechanism, allowing for the effective control of user access during different time intervals. This enhancement makes the system more adaptable to complex security scenarios.

2.8. Time-Bound

Time-bound mechanisms are essential for controlling data access in encryption systems, providing better security and flexibility. These mechanisms limit access to data for a specified period, after which permissions are automatically revoked. This prevents long-term exposure to unauthorized users [27]. In cloud computing, time-bound mechanisms ensure that data are accessible only when needed, reducing the risk of data leakage. They also support compliance with legal and regulatory requirements, such as retention periods for medical or financial data, by automating access control and data deletion [28]. Additionally, time-bound mechanisms improve key management by assigning expiration periods to keys. This ensures that keys are only valid for specific timeframes, enhancing security in long-term operations [29].

The demand for time-bound mechanisms stems from several key factors. (1) Data often only needs to be accessible for specific time periods. In cloud storage, for example, access must be dynamically controlled based on user needs and authorization periods. Time-bound mechanisms adjust data availability accordingly [27]. (2) Prolonged data access increases security risks. Time-bound mechanisms reduce this risk by allowing access only for necessary periods and automatically revoking it afterward, minimizing the chance of data leakage. (3) In distributed applications like the Internet of Things (IoT) and edge computing, dynamic data management is critical. Time-bound mechanisms ensure that devices can access data only during authorized periods, tailored to their specific needs and contexts. Key applications include: (1) dynamic key management, where keys are valid for limited periods and automatically updated or expired to enhance security [28]; (2) data access control for sensitive information which ensures that data are accessible only within a specific timeframe, preventing unauthorized access, even if credentials are later compromised; (3) hierarchical security systems that use time-bound mechanisms to dynamically adjust access permissions for different user levels, commonly used in high-security fields such as finance and healthcare.

The implementation of time-bound mechanisms uses various techniques to improve system security and flexibility by enforcing time limits. A common method is hierarchical key assignment, where users receive keys based on their access level and specific timeframes. These keys become invalid after the designated time, ensuring that data can only be encrypted or decrypted within authorized periods [27]. Another method is dynamic key updates, where encryption keys are automatically refreshed at regular intervals. This ensures that a new key is used for each time period. This approach limits the damage caused by key compromise, as only data from one time period is affected [28]. Key revocation is another important feature, where keys are automatically revoked after expiration to prevent unauthorized access over time [29]. Several studies have explored time-bound mechanisms. For example, Tseng et al. proposed the Time-Bound Hierarchical Key Assignment (TBHKA) scheme, which combines RSA and pairing-based encryption to improve the efficiency and security of key assignments in hierarchical systems [30]. Time-bound mechanisms are particularly useful in cloud computing, IoT, and regulatory compliance, where long-term data protection and dynamic key management are essential.

Chien proposed a time-bound method in 2018, primarily designed to provide a time-restricted access mechanism for IoT services [31]. This approach allows users to access specific resources only within a designated time frame. The concept of time-bound refers to the restriction that users can only access resources during the valid period authorized by the server. When a user requests access permission from the server, the server sets a permissible time period for the user. Once set, the user can access the service within this authorized period. However, if the user attempts to access the service outside the authorized timeframe, the server will deny the request. This time-bound mechanism is applicable in various scenarios, such as online courses, music subscriptions, and membership-based services.

As shown in Figure 2, the valid period is set from 01/10 to 01/18. Here, $t_1=10$ and $t_2=18$. If a user attempts to access the service on 01/08, which falls outside the valid range, the server calculates the identity authentication token based on the requested time. Since 01/08 is earlier than the valid start time, the number of hash computations would be negative, resulting in a failed computation. Consequently, the server denies access. On the other hand, if the user tries to access the service on 01/12, which is within the valid period, the server calculates a positive value for the authentication token. This positive result confirms that the user’s request falls within the authorized timeframe, and the server grants access.

3. Research Process

This chapter primarily demonstrates the proposed method of this study through a contextual scenario. The demonstration is divided into two aspects: data types and scenario processes.

There are two main types of data: the game’s source code and Personal Health Records (PHR). The VR-Based Health promotion game includes a total of 15 games, categorized into three major types: Cerebrovascular Accident (CVA) prevention, health promotion, and osteoporosis prevention. Each category contains five games. These data are encrypted using a time-bound method. Originally, time-bound encryption restricted access based on time intervals. Here, we modify the method to restrict access by game segments.

The game segments are as follows:

• Cerebrovascular Accident prevention—Games 1 to 5 defined as G1;

• Health promotion—Games 6 to 10 defined as G2;

• Osteoporosis prevention—Games 11 to 15 defined as G3.

Subscriptions can be customized to specific intervals, e.g., Site’s subscribing to games 3 to 10. In such cases, the algorithm’s $t{n}_1$ and $t{n}_2$ represent the start and end points of the subscribed segment.

PHRs are divided into three categories: stroke-type, fitness-type, and osteoporosis-type, each corresponding to specific KP-ABE attributes. Regarding case manager roles, they are classified into four types:

1. Physical Examination Center Site Case Manager;

2. Cerebrovascular Accident (CVA) Site Case Manager;

3. Health promotion Site Case Manager;

4. Osteoporosis Site Case Manager.

Their access rights to PHRs are outlined in Table 2. The Health Check Center Case Manager has access to all PHRs, while the CVA, Health promotion, and Osteoporosis Case Managers can only access PHRs within their respective specialties. For example, in Figure 3, SCM1 can access CVA-related data, which includes performance records for game G1. Some sites may subscribe to more than one type of game, meaning that a single individual’s PHR could contain performance data from games G1, G2, or G3.

The process begins with the initialization by the trusted authority (TA). The Site then registers with the TA, specifying the game segments ($t{n}_1,t{n}_2$) and the subscription period ($t{d}_1,t{d}_2$), which $t{d}_1,t{d}_2$ is the start and end time of Site subscription. In Figure 4, we can see that the first <1,5> for Site1 represents the subscribed game segments, $t{n}_1,t{n}_2$. It is worth noting that Site5 can subscribe to all games, meaning cross-segment subscription is allowed. Additionally, the <10,18> for Site1 indicates the subscription period $t{d}_1,t{d}_2$, which means that the subscription starts from the 10th day of the year and ends on the 18th day. The TA generates the corresponding Time-bound Key (TK) for the Site and sends the subscription duration to the Content Service Provider (CSP). Based on the subscription requirements, the CSP generates the encrypted game source code. Subsequently, the TA produces the final encrypted file (CC) and delivers it to the Site. Upon receiving the CC, the Site uses the TK provided by the TA to decrypt the game. When launching the game, the system checks whether the current date falls within the subscribed period ($t{d}_1,t{d}_2$). If valid, the game is successfully launched.

After gameplay, a Personal Health Record (PHR) is generated and sent to the TA for centralized management. When a Site Case Manager needs to access the PHR, they request it from the TA. Since the Case Manager’s role and permissions were established during the registration phase, the TA generates a corresponding Secret Key (SK) based on the Case Manager’s attributes. Using this key, the Case Manager can decrypt the PHR and review the health data of the game users at the Site. This process not only ensures the security of digital rights but also achieves fine-grained access control and data protection.

The process begins with a Site subscribing to VR-based health promotion games. The Site sends its subscription details to the trusted authority (TA), including the start of the game segment ($t{n}_1$), the end of the game segment ($t{n}_2$), the subscription start time ($t{d}_1$), and the subscription end time ($t{d}_2$). The TA then generates a Time-bound Key (TK), which is specifically associated with the subscribed game segments ($t{n}_1$, $t{n}_2$) and returns it to the Site. This step is completed during the Site Registration Phase.

Subsequently, the TA forwards the Site’s request to the Content Service Provider (CSP). The CSP encrypts the games using the specified subscription start and end times ($t{d}_1,t{d}_2$) and organizes the games into 15 folders, one for each game. These encrypted games are then sent back to the TA. The TA applies time-bound encryption to the subscribed game segments based on the Site’s request and sends the encrypted game files (CT) to the Site. This step corresponds to the Game Encryption Phase.

Next, the Site decrypts the games using its time-bound key (TK). This step aligns with the Site Game Decryption Phase. When a game is launched, it performs a self-verification to ensure that the gameplay date falls within the subscribed period, ensuring that the game runs only within the authorized time frame. This verification process corresponds to the Site Game Verification Phase.

After gameplay, a Personal Health Record (PHR) is generated and uploaded to the TA. The TA encrypts the PHR based on its attributes, as shown in Figure 3. The encryption algorithm for this step is detailed in the Site PHR Encryption Phase. Subsequently, Site Case Managers (SCM) register their attributes and receive their corresponding private keys, as described in the Site Case Manager Registration Phase.

When an SCM needs to access a PHR, they send a request to the TA. The TA then sends the corresponding ciphertext (CT) to the SCM, who uses their registered private key to decrypt and read the PHR. This process corresponds to the Site Case Manager Decryption Phase.

Finally, to ensure data security, the system updates the PHR ciphertexts and SCM private keys monthly. The updated ciphertexts are broadcasted, and the updated private keys are sent to the SCM. This mechanism ensures that PHR data remains secure even in the event of key exposure.

4. Methodology

4.1. System Framework of a Game-Based Personal Health Promotion System

There are four main roles: The Trusted Authority (TA), the Content Service Provider (CSP), Site Case Manager (SCM), and the Site. The trusted authority generates public parameters (Pub_info), private parameters (Priv_info), and the public (Pub_phrinfo) and private parameters (Priv_phrinfo). The Site subscribes to the desired game intervals and subscription periods based on its needs. Subsequently, the TA forwards the Site’s request tn to the CSP, which generates the corresponding subscription time frame and returns the game data to the TA. The TA then generates the TK for the subscription interval based on the Site’s request and broadcasts the encrypted game content (CC) to the Site. The Site uses the TK to decrypt and play the game. During game activation, a date verification is performed to ensure that the gameplay falls within the subscription period. Once confirmed, the game successfully launches.

After gameplay, corresponding PHRs (Personal Health Records) are generated, with different games producing various PHR fields. The PHRs are uploaded to the TA, which encrypts them using the specific fields and generates CT (Ciphertexts) for storage. The Site Case Manager (SCM) can request access to PHRs based on medical needs and their role. The TA generates an SK (Secret Key) based on the SCM’s registered attributes, enabling the SCM to decrypt the CT and access relevant data fields.

The game employs time-bound encryption, leveraging time-bound features to ensure access to specific intervals, thereby securing digital rights management. This ensures that the Site can access only authorized game types. Game verification ensures operations are restricted to the Site’s subscription period, preventing unauthorized game launches. The generated PHRs are encrypted using KP-ABE (Key-Policy Attribute-Based Encryption). SCM can register attributes and obtain the secret key (SK). Decryption is based on attributes, enabling fine-grained access control. Key insulation updates the ciphertext (CT) and SK to protect against key exposure. Figure 5 shows the system framework.

4.2. A Game-Based Personal Health Promotion System with Time-Bound Secure Access Control

This method introduces the key insulated key strategy attribute encryption method with a time-bounded mechanism. Table 3 shows notations of the proposed scheme. It is mainly divided into seven phases, namely the system initialization phase, the Site registration phase, the Site Case Manager registration phase, the game encryption phase, the Site PHR encryption phase, the game decryption phase, the Site Case Manager decryption phase, the game verification phase, the key update phase and the ciphertext update phase. The system initialization phase mainly describes the initial settings of the system, the Site registration phase mainly generates a time-bound key according to the application time, the Site Case Manager registration phase generates a KP-ABE secret key according to the Site Case Manager’s attributes, the game encryption phase mainly encrypts the game itself and broadcasts it, the Site PHR encryption phase encrypts PHR by the attributes, the Site Case Manager decryption phase decrypts the CT by the corresponding attribute secret key, and the game decryption phase mainly decrypts the game once. The Site game verification phase mainly verifies and updates the game based on time. The key update phase aims to update the key to achieve the purpose of key insulation. The ciphertext update phase updates the ciphertext to achieve the purpose of key insulation.

4.3. The System Initialization Phase

The system initialization phase is set by the trusted authority, and public parameters and secret parameters are generated through the following steps, the procedure is shown in Algorithm 1:

• Step 1.. The trusted authority selects a large prime number $q$ and sets GF(q) as the finite field with order $q$. It then determines the elliptic curve equation $E:y^2=x^3+ax+b\mathrm{mod}q,$ where $a$ $\mathrm{and}b$ are elements of GF(q) and 4$a^3+27b^2\ne 0$.

• Step 2.. The trusted authority generates an additive group $G_1\mathrm{and}\mathrm{a}\mathrm{multiplicative}\mathrm{group}{G}_T,$ ensuring that the bilinear map function is $e:G_1\times G_1\to G_T$.

• Step 3.. The trusted authority sets $g$ as the generator of $G_1$.

• Step 4.. The trusted authority defines a one-way hash function $H:\{0,1{\}}^{\ast }\in G_T$.

• Step 5.. The trusted authority computes the public key $Y=e{\left(g,g\right)}^a$.

• Step 6.. The trusted authority generates a random secret value $x_s,x_e\in Z_p$ which serves as a time-bound secret key.

• Step 7.. The trusted authority sets system attribute domain $u_s=\left\{at{t}_1,at{t}_2,\dots ,at{t}_n\right\},$ where n is the number of attributes in attribute set $u_s.$

• Step 8.. The trusted authority selects the corresponding random points $\left\{Q_1,Q_2,\dots ,Q_n\right\}{\in }_R{G}_1$ according to the attributes $u_s$.

• Step 9.. The trusted authority selects the corresponding secret random values $t_1,t_2,$…$,t_n$ ∈ $Z_p^{\ast }$ based on the attributes $u_s$.

• Step 10.. The trusted authority randomly selects values $a,v,y{\in }_R{Z}_p$.

• Step 11.. The trusted authority calculates $P_{cs}$ = v$g$.

• Step 12.. The trusted authority calculates $Y^{\prime }$ = $ag$.

• Step 13.. The trusted authority calculates $T_i$ = $t_{i}g$ for i = 1, 2, …, n.

• Step 14.. The trusted authority outputs Pub_phrinfo = {$P_{cs},$ Y′, $Q_1,$…$,Q_n,T_1$,…$,T_n$} and Priv_phrinfo = {$a,v,y,t_1$,…$,t_n$}.

• Step 15.. The trusted authority outputs Pub_info = {$g,$Y} and Priv_info = {$x_s,x_e$}.

Algorithm 1 Sys_Int: System initialization

Trusted Authority (TA)

1.  $E:y^2=x^3+ax+b\mathrm{mod}q$2.  $e:G_1\times G_1\to G_T$3.  $g\in G_1$4.  $H:\{0,1{\}}^{\ast }\in G_T$5.  $Y=e{\left(g,g\right)}^a$6.  $x_s,x_e\in Z_p$7.  $u_s=\left\{at{t}_1,at{t}_2,\dots ,at{t}_n\right\}$8.  $\left\{Q_1,Q_2,\dots ,Q_n\right\}{\in }_R{G}_1$9.  $t_1,t_2,$…$,t_n$ ∈ $Z_p^{\ast }$10. $a,v,y{\in }_R{Z}_p$.11. $P_{cs}$ = v$g$.12. $Y^{\prime }$ = $ag$.13. $T_i$ = $t_{i}g$ for i = 1, 2, …, n.14. Pub_info = {$g,$Y}   Priv_info = {$x_s,x_e$}15. Pub_phrinfo = {$P_{cs}$, Y′,$Q_1$, …,$Q_n$,$T_1$, …,$T_n$}   Priv_phrinfo = {$a,v,y,t_1$,…$,t_n$}

4.4. Site Registration Phase

In the Site registration phase, the Site subscribes to the TA, providing the subscribed game segment $t{n}_1,t{n}_2$ and subscription period $t{d}_1,t{d}_2$. The trusted authority performs the following steps to generate the corresponding private key ${\mathit{SK}}_i$ and time-bound keys $T{K}_1\mathrm{and}T{K}_2$. The steps are as follows and the procedure is shown in Algorithm 2:

• Step 1.. The Site sends the requirements ($t_1,t_2,t{n}_1,t{n}_2$) to the trusted authority.

• Step 2.. The trusted authority generates $T{K}_1=H^{t{n}_1-1}\left(x_s\right),T{K}_2=H^{15-t{n}_2}\left(x_e\right)$.

• Step 3.. The trusted authority sends $T{K}_1,T{K}_2$ to the site.

Algorithm 2 Site_Reg: Site registration with secret key generation
Site	Trusted Authority (TA)
1. ($t_1,t_2,t{n}_1,t{n}_2$)
	2. $T{K}_1=H^{t{n}_1-1}\left(x_s\right),T{K}_2=H^{15-t{n}_2}\left(x_e\right)$
	3. $T{K}_1,T{K}_2$

4.5. Site Case Manager Registration Phase

In the Site Case Manager (SCM) registration phase, the SCM requests the private key ${\mathit{SK}}_i$, from the trusted authority based on its attributes. Additionally, the Site sends the corresponding attributes to TA. The trusted authority performs the following steps to generate the corresponding private key ${\mathit{SK}}_i$. The steps are as follows and the procedure is shown in Algorithm 3:

• Step 1.. The SCM sends the attributes ($u_x$) to the trusted authority.

• Step 2.. The TA randomly generates $r_1,$…$,r_l{\in }_R{Z}_q,$ where l is the number of attributes of the SCM.

• Step 3.. TA calculates $L_j$ = ($\left(p_R\left(0\right)·{t_j}^{-1}\right)g$ + $Y^{\prime }$), where j = 1$,$ 2$,$…$,$ l.

• Step 4.. TA calculates $K_j$=$T_j$ − $r_l{Q}_j$.

• Step 5.. TA calculates $K_{j,0}$ = $r_j\left(P_{cs}+H_1\left(T_j,0\right)g\right)$.

• Step 6.. TA sends the private key ${\mathrm{SK}}_i$ = <$L_j,K_j,$ $K_{j,0}$> to the Site Case Manager, where j = 1$,$2$,$…$,$l.

Algorithm 3 SCM_Reg: Site Case Manager registration with secret key generation
Site Case Manager (SCM)	Trusted Authority (TA)
1. $u_x$
	2. $r_1$,…$,r_l{\in }_R{Z}_q$
	3. $L_j$ = ($\left(p_R\left(0\right)·{t_j}^{-1}\right)g$ + $Y^{\prime }$)
	4. $K_j$=$T_j$ − $r_l{Q}_j$
	5. $K_{j,0}$ = $r_j\left(P_{cs}+H_1\left(T_j,0\right)g\right)$
	6. ${\mathrm{SK}}_i$ = <$L_j,K_j,$ $K_{j,0}$>

4.6. Game Encryption Phase

In the game encryption phase, the TA sends the subscribe information $t{d}_1,t{d}_2,t{n}_1,t{n}_2$ to the content service provider (CSP). CSP generates $M_1$ and send it to the TA. The TA generates ciphertext of the game through the following steps; the procedure is shown in Algorithm 4:

• Step 1.. The content service provider confirms the attribute values for the TA as ($t{d}_1,t{d}_2,t{n}_1,t{n}_2$).

• Step 2.. The content service provider calculates the game $M_1$ = $H^n(Game||Date)$ and send it to TA.$n=t{d}_2-t{d}_1$.

• Step 3.. The trusted authority then calculates $TK=$ $H(H^{t_1-1}\left(x_s\right)||H^{365-t_2}\left(x_{\mathrm{e}}\right))$.

• Step 4.. The trusted authority calculates $CC=$ $E_{TK}\left(CT\right)$.

• Step 5.. The trusted authority outputs $CC$.

Algorithm 4 Game_Enc: Game encryption
Content Service Provider (CSP)	Trusted Authority (TA)
	1. $t{d}_1,t{d}_2,t{n}_1,t{n}_2$
2. $M_1=H^n(Game||Date)$
	3. $TK=$ $H(H^{t{n}_1-1}\left(x_s\right)||H^{15-t{n}_2}\left(x_{\mathrm{e}}\right))$
	4. $CC=$ $E_{TK}\left(M_1\right)$
	5. output $CC$

4.7. Site PHR Encryption

In the Site PHR encryption phase, the Site sends PHR to the TA. The TA generates CT by the following steps, the procedure is shown in Algorithm 5:

• Step 1.. The Site sends PHR to TA.

• Step 2.. The TA randomly generates a value $s\in Z_p$.

• Step 3.. The TA calculates the ciphertext C = $M_2{Y}^s$.

• Step 4.. The TA calculates $C^{\prime }=\mathrm{s}{t}_{u}g,$ where u is the number of game attributes.

• Step 5.. The TA calculates $C^{″}=s{Y}^{\prime }$.

• Step 6.. The TA generates $E{K}_{u,0}$= ($a·{\left(v+H_1\left(T_u,0\right)\right)}^{-1}$) $Q_u$ for all$u\in$l$,$ where l is the set of attributes of PHR.

• Step 7.. The TA calculates $C_{i,u,0}$=$s(E{K}_{u,0}$)$,$ where $i$ is the game identifier.

• Step 8.. The TA outputs CT = < $C,C^{\prime },C^{″},C_{i,u,0};\forall u\in I$> and stores it.

Algorithm 5 PHR_Enc: Site PHR Encryption
Site	Trusted Authority (TA)
1. $M_2=\mathrm{PHR}$
	2. $s\in Z_p$
	3. $C=M_2{Y}^s$
	4. $C^{\prime }=\mathrm{s}{t}_{u}g$
	5. $C^{″}=s{Y}^{\prime }$
	6. $E{K}_{u,0}$= ($a·{\left(v+H_1\left(T_u,0\right)\right)}^{-1}$) $Q_u$
	7. $C_{i,u,0}$=$s(E{K}_{u,0}$)
	8. CT =< $C$,$C^{\prime }$,$C^{″}$,$C_{i,u,0};\forall u\in I$>

4.8. Site Game Decryption Phase

In the Site game decryption phase, the Site receives the game ciphertext from the trusted authority and, using the time-bound key $T{K}_1,T{K}_2,$ uses CC to restore the game ciphertext to plaintext through the following steps; the procedure is shown in Algorithm 6:

• Step 1.. The trusted authority sends ($CC$) to the site.

• Step 2.. The Site calculates $H^{t-t_1}\left(T{K}_1\right) \mathrm{and}{H}^{t_2-t}\left(T{K}_2\right)$.

• Step 3.. The Site calculates $T{K}^{\prime }=$ $H(H^{t-t_1}\left(T{K}_1\right)||H^{t_2-t}\left(T{K}_2\right))$.

• Step 4.. The Site calculates CT = $D_{T{K}^{\prime }}\left(CC\right)$.

• Step 5.. The Site obtains $M_1\mathrm{from}CC$.

Algorithm 6 Game_Dec: Site game decryption
Trusted Authority (TA)	Site
1. $CC$
	2. $H^{tn-t{n}_1}\left(T{K}_1\right)\mathrm{and}{H}^{t{n}_2-tn}\left(T{K}_2\right)$
	3. $T{K}^{\prime }=$ $H(H^{tn-t{n}_1}\left(T{K}_1\right)||H^{t{n}_2-tn}\left(T{K}_2\right))$
	4. $M_1$ = $D_{T{K}^{\prime }}\left(CC\right)$
	5. output $M_1$

4.9. Site Game Verification Phase

During the verification phase of the Site game, every time the Site game is opened, it will initiate a contest. The Site will verify the game file to make sure the game is activated in a valid time. The procedure is shown in Algorithm 7. The steps are as follows:

• Step 1.. The Site takes $M_1$ as input.

• Step 2.. The Site calculates $H^{t{d}_2-td}\left(H^{td-t{d}_1}\left(Game\left|\right|Date\right)\right),$ where $t$ is the current time, $t{d}_1$ is the start of the subscribe time, and $t{d}_2$ is the end of the subscribe time.

• Step 3.. The Site verifies where

  $H^{t{d}_2-td}\left(H^{td-t{d}_1}\left(Game\left|\right|Date\right)\right)\mathrm{matches}{H}^n\left(Game\left|\right|Date\right).$

• Step 4.. If the verification is successful, the Site successfully starts the game.

• Step 5.. If$t{d}_2-td$ = 0, the game cannot be started.

Algorithm 7 Game_Verif: Game Verification

Site

1.  Perform Site Decryption Phase obtain $M_1$2.  $H^{t{d}_2-td}\left(H^{td-t{d}_1}\left(Game\left|\right|Date\right)\right)$3.  $H^{t{d}_2-td}\left(H^{td-t{d}_1}\left(Game\left|\right|Date\right)\right)·H^n\left(Game\left|\right|Date\right)$4.  Start game5.  If $t{d}_2-td$ = 0⊥

4.10. Site Case Manager Decryption Phase

In Site Case Manager decryption phase, TA will send the CT due to Site Case Manager’s request. The SCM will decrypt the CT their own SK which related to their own attributes. The procedure is shown in Algorithm 8. The steps are as follows:

• Step 1.. The trusted authority sends ($CT$) to the site.

• Step 2.. The Site Case Manager uses their private key and the ciphertext to compute

  $e(g,g{)}^{as}modp=(\mathrm{e}(C^{\prime },L_j)·{{\displaystyle \prod }}_{x\in I}\left(e\left(K_j,C^{″}\right)e\left(K_{j,0},C_{i,u,0}\right)\right){)}^{-1}modp.$

• Step 3.. C$·e(g,g{)}^{-ys}$mod p = $M_{2}e(g,g{)}^{ys}\mathrm{mod}p·e(g,g{)}^{-ys}modp=M_2$

• Step 4.. The Site Case Manager obtains the plaintext $M_2$.

Algorithm 8 SCM_Dec: Site Case Manager decrypt CT into PHR
Trusted Authority (TA)	Site
1. CT
	2. $e(g,g{)}^{as}modp$   $=(\mathrm{e}(C^{\prime },L_j)·{{\displaystyle \prod }}_{x\in I}\left(e\left(K_j,C^{″}\right)e\left(K_{j,0},C_{i,u,0}\right)\right){)}^{-1}modp$
	3. C$·e(g,g{)}^{-ys}modp$      $=M_{2}e(g,g{)}^{ys}modp·e(g,g{)}^{-ys}modp=M_2$
	4. output $M_2$

4.11. Ciphertext Update Phase

In the ciphertext update phase, the goal is to achieve key isolation through the update of ciphertext. This is performed by the trusted authority. After the update, the corresponding keys must be used for decryption. The procedure is shown in Algorithm 9. The steps are defined as follows:

• Step 1.. The trusted authority calculates $H{K}_{u,1}$ = ($v+H_1$($T_u,$ 0)) $·$ ${(v+H_1\left(T_u,1\right))}^{-1}$.

• Step 2.. The trusted authority calculates $C_{i,u,1}$ = $C_{i,u,0}·$$H{K}_{u,1}$.

• Step 3.. The trusted authority outputs CT_new = < $\mathrm{C},$C′, C″, $C_{i,u,1};\forall u\in I$>.

Algorithm 9 CT_Upd: Cyphertext update and broadcast

Trusted Authority (TA)

1.  $H{K}_{u,1}$= ($v+H_1$($T_u$, 0)) $·$ ${(v+H_1\left(T_u,1\right))}^{-1}$.2.  $C_{i,u,1}$= $C_{i,u,0}·$ $H{K}_{u,1}$.3.  CTnew = <$\mathrm{C}$,C′,C″,$C_{i,u,1};\forall u\in I$>.

4.12. Key Update Phase

In the key update phase, the goal is to achieve key isolation through the update of keys. This is conducted by the trusted authority. After the update, the updated keys are used to decrypt the corresponding ciphertext. The procedure is shown in Algorithm 10. The steps are as follows:

• Step 1.. The trusted authority calculates $H{K}_{j,1}$= ($v+H_1$($T_j,$ 0))$·$ ${(v+H_1\left(T_j,1\right))}^{-1}$.

• Step 2.. The trusted authority calculates $K_{j,1}$ = $K_{j,0}·H{K_{j,1}}^{-1}$.

• Step 3.. The trusted authority outputs the updated secret key ${\mathrm{SK}}_i$ = <$L_j,K_j,$ $K_{j,1}$>.

• Step 4.. Sends the updated private key ${\mathrm{SK}}_i$ to the site.

Algorithm 10 Key_Upd: Site Case Manager’s secret key update

Trusted Authority (TA)

1. $H{K}_{j,1}$= ($v+H_1$($T_j$, 0))$·$ ${(v+H_1\left(T_j,1\right))}^{-1}$2. $K_{j,1}$ = $K_{j,0}·H{K_{j,1}}^{-1}$3. ${\mathrm{SK}}_i$ = <$L_j,K_j,$ $K_{j,1}$>4. The trusted authority outputs the updated private key ${\mathrm{SK}}_i$ = <$L_j,K_j,$ $K_{j,1}$> and sends it to the site.

4.13. Security Analysis

This chapter focuses on difficulty, correctness, confidentiality, key security, and time-bound mechanisms. It also provides a detailed analysis of the security of this method in each subsection.

4.14. Elliptic Curve Discrete Logarithm Problem

Elliptic Curve Cryptography (ECC) defines the Elliptic Curve Discrete Logarithm Problem (ECDLP). Given an elliptic curve ECC, a finite field $GF\left(q\right)$, and a random integer k, two different points $q$ and $Q$ are selected on the elliptic curve $E$, such that $Q=kq$. If $k$ and $q$ are known, $Q$ can be easily computed. However, when only $q$ and $Q$ are known, it is difficult to find $k$ that satisfies $Q=kq$.

4.15. Correctness

The correctness of the Site Decryption Phase algorithm during the ciphertext decryption phase is verified as follows.

Verification of the correctness of the parameter $e(g,g{)}^{\alpha s}$:

$\begin{array}{l}e(g,g{)}^{\alpha s}=\mathrm{e}(C^{\prime },L_i)·{({{\displaystyle \prod }}_{x\in I}\left(e\left(K_i,C^{″}\right)e\left(K_{i,0},C_{i,x,0}\right)\right))}^{-1}\\ =e(\mathrm{s}{t}_{i}g,\left(p_R\left(0\right)·{t_i}^{-1}\right)g+ag)·({{\displaystyle \prod }}_{i\in I}{\left(e\left({\mathrm{t}}_{i}g-r_i{Q}_i),\mathrm{s}ag\right)e\left(r_i\left(vg+H_1\left(T_i,0\right)g\right),(\mathrm{s}a·({\left(v+H_1(T_x,0\left)\right)Q_x\right)}^{-1}\right)\right)}^{-1}\\ ={(e(g,g)}^{as{t}_i}e\left(g,g{)}^{p_R\left(0\right)s}\right)·{(e\left(g,g\right))}^{-sa{t}_i}\\ ={(e(g,g)}^{as{t}_i}e\left(g,g{)}^{p_R\left(0\right)s}\right)·{(e\left(g,g\right))}^{-as{t}_i}\\ =e(g,g{)}^{ys}\end{array}$

Verification of the correctness of the decrypted plaintext:

$\begin{array}{l}C·e(g,g{)}^{-ys}\\ =Me(g,g{)}^{ys}\mathrm{mod}p·e(g,g{)}^{-ys}·\mathrm{mod}p\\ =M\end{array}$

4.16. Confidentiality

Confidentiality is divided into two parts: the risk of external attackers (Outsider Risk) and the risk of internal attackers (Insider Risk). External risk refers to attackers trying to access unauthorized data. Internal risk involves collusion attacks or attempts to steal data from within. These risks are discussed as follows.

• Outsider Risk

If an attacker wants to obtain the message M, they must first acquire the ciphertext CT and decrypt it. From the Enc algorithm, to decrypt, the attacker must either obtain or forge the following parameters: $C,C^{\prime },C^{″},C_{i,u,0}$. The details of these parameters are explained below:

• To break $\mathrm{C}=M·Y^s$, the attacker must obtain $Y^s$, where s is the secret parameter used to generate the ciphertext, and $Y=e{\left(g,g\right)}^a$, which is protected by the Elliptic Curve Discrete Logarithm Problem (ECDLP), making it impossible to directly obtain the secret parameter $s$;

• To break $C^{\prime }=\mathrm{s}{t}_{u}g$, the attacker must obtain the secret parameters $s$ and $t_u$, both of which are protected by the ECDLP, making it impossible to obtain the secret parameters $s$ and $t_u$;

• To break $C^{″}=s{Y}^{\prime }$, the attacker must obtain the public parameter $Y^{\prime }$ and the secret parameter $s$, where $Y^{\prime }=a·g$ and a is the secret parameter of the system, protected by the ECDLP, making it impossible to obtain both the secret parameter $a$ of the system and the secret parameter $s$ of the ciphertext;

• To break $C_{i,u,0}$, one must obtain $s(a·{\left(v+H_1\left(T_u,0\right)\right)}^{-1}$) $Q_u;\forall u\in I$. Since $H_1\left(T_u,0\right)$ is protected by the hash function, it is impossible to obtain the parameters $s,a\mathrm{and}v$.

Based on the above analysis, it can be concluded that the message M is protected by ECDLP and hash functions in this method.

• Insider Risk

If an attacker attempts to obtain unauthorized ciphertext through an internal attack, they will primarily attempt to forge the attribute private key $t_u$ and the update key $H{K}_{j,1}$. These risks are discussed below:

• If an attacker attempts to recover $e(g,g{)}^{ys}$, they must possess the corresponding attribute private key $t_u$. Without it, they cannot derive the unique solution for the linear combination $y=p_R\left(0\right)$, thus making it impossible to decrypt the message M.

• If the private key ${SK}_i$ has not been updated, it will be impossible to decrypt the updated ciphertext $C_{i,u,1}$ due to the lack of the updated parameter $K_{j,0}$ in the private key, which is required to calculate the decryption parameter ${(e\left(g,g\right))}^{-as{t}_j}$.

Based on the above analysis, it can be concluded that these internal attacks of the method are protected by elliptic curve cryptography and key insulation.

4.17. Key Security

Assume that an illegal attacker wants to obtain the secret parameter a of the system through the public parameter $Y^{\prime }$ = $ag$. Because $Y^{\prime }$ = $ag$ is protected by solving the elliptic curve discrete logarithm problem, the secret parameter $a$ cannot be obtained. If an attacker wants to use the public parameter $P_{cs}$= v$g$ to find the secret key v of the trusting authority, he will be unable to find the secret key v of the trusting authority because $P_{cs}$= v$g$ is protected by the solution of the elliptic curve discrete logarithm problem.

If an attacker tries to combine private attribute keys from multiple users to match the attribute set required by the game, they will fail. Each attribute private key, $K_{i,0}$ = $r_i\left(P_{cs}+H_1\left(T_i,0\right)g\right)$, includes a unique parameter $\left(r_i\right)$, This parameter is generated by the trusted authority (TA) and remains a secret. Therefore, attackers cannot use collusion attacks to decrypt the ciphertext.

If an attacker steals an attribute private key from someone, he will not be able to access ciphertext in different periods due to the insulating nature of the key; that is, the attribute private key $K_{i,0}$ = $r_i\left(P_{cs}+H_1\left(T_i,0\right)g\right)$ will be updated to $K_{i,0}$= $r_i\left(P_{cs}+H_1\left(T_i,0\right)g\right)K_{i,0}$= $r_i\left(P_{cs}+H_1\left(T_i,1\right)g\right)$, and the parameters in the ciphertext will also be updated to $\mathrm{s}a/v+H_1(T_x,1\left)\right)Q_x$, so if the attacker obtains the attribute private key $K_{i,0}$= $r_i\left(P_{cs}+H_1\left(T_i,0\right)g\right)$, he still cannot decrypt the ciphertext.

4.18. Time-Bound Mechanism

The time-bound mechanism allows users to access resources only within the server’s authorized validity period. When a user requests access permission, the server sets a specific time period. Users can access the service only during this authorized period. Therefore, if an attacker wants to forge the time, he will not be able to forge the key because there is no $x_s$ and $x_e$ in $T{K}_1=H^n\left(x_s\right),T{K}_2=H^n\left(x_e\right)$.

If the $T{K}_1$ and $T{K}_2$ obtained by the attacker can only access this broadcast, the next broadcast will be generated by generating a new set of $T{K}_1$ and $T{K}_2$ that are different $x_s$ and $x_e$ Different values are output, and the time-bounded ciphertext of different broadcasts cannot be accessed again.

5. System Demonstration

The system architecture is shown in Figure 6. We assume that each Site will have both users and healthcare personnel who access the system to play games. All individuals use the system within the Site, and the system administrator at the Site can manage related user data. This method primarily protects the games themselves, which represent the entire system.

The trusted authority (TA) generates public and secret parameters, and each Site registers based on its required game requests. The TA then generates a private key for the Site, creating a hash value from the game identifier and date. It sets up a hash chain based on the Site’s requested time, encrypting this hash value into ciphertext. Using a time-bound mechanism, the TA generates a time-bound key and time-bound ciphertext for the specified time interval requested by the Site. This time-bound ciphertext contains the previously encrypted hash value. The TA sends the time-bound key to the Site and broadcasts the time-bound ciphertext. The Site then decrypts the time-bound ciphertext using the time-bound key; if the requested time does not match, the Site cannot decrypt it. After decrypting the time-bound ciphertext, the Site uses its private key based on its attributes to decrypt the ciphertext, obtaining the hashed game identifier and date. Before launching the game, the Site verifies by calculating the hash chain value according to the current time. If the verification is successful, the game can be accessed.

Figure 7 primarily displays the Cerebrovascular Accident gameplay, featuring various game categories such as CVA prevention and dementia prevention. This figure illustrates six different game interfaces, each with a distinct activity. The activities include (from left to right, top to bottom): avoiding falling rocks, catching gold, digging Carrots, chasing and jumping, catching butterflies, and farming soil. The titles for each game are written in Chinese characters. Sites can register and subscribe to specific games based on their needs; if a game is not subscribed to, it cannot be accessed. Additionally, if the subscription period has expired, access to the game will be restricted due to the time-bound mechanism.

6. Performance Analysis

According to [32], the time for performing an exponentiation operation ≈ 522 ms, and in [33], usually the pairing operation cost five times that of an exponentiation operation ≈ 2610 ms; $\left|p\right|$ is assumed as 1024 bits.

Hong et al. [26] proposed a CP-ABE scheme with key insulation, while He et al. [21] proposed an efficient KP-ABE method. As compared with [26], both our method and theirs include key insulation, but Hong et al.’s scheme [26] uses CP-ABE, while our method uses KP-ABE. Additionally, Hong et al.’s scheme [26] does not use time-bound encryption. In contrast, our system integrates time-bound encryption with KP-ABE to support both games and PHR. The time-bound process only requires hash computation, which has minimal computational cost compared to exponential and pairing operations. Therefore, its overhead is considered negligible

As compared with He et al.’s scheme [21], our method introduces key insulation and incorporates time-bound encryption. In addition to solving the key leakage problem, our approach establishes a dual system for games and PHR, with distinct access control mechanisms for each. Furthermore, our method demonstrates better performance in terms of computational efficiency.

In Table 4, we analyze the number of exponentiation and pairing operations required in the calculation method, where n represents the number of attributes. This allows us to compare the computational efficiency of our method with Hong et al.’s scheme [26] and He et al.’s scheme [21]. For example, in Hong et al.’s encryption scheme [26], 3n attribute-related exponentiation operations and three general exponentiation operations are performed. These computations are used to estimate the total execution time. This study adopts the computation speed proposed in Banerjee et al.’s scheme [32], calculating the number of exponential operations, pairing operations, and attributes. According to reference [33], pairing operations take abou34 five times longer than exponential operations. An exponential operation requires 522 ms, while pairing operations take approximately 2610 ms. n remains representative of the number of attributes. Regarding the sizes of private keys and ciphertexts, based on the Site registration and encryption phases, the private key size for a single user is (2n + 1)$\left|p\right|$, and the ciphertext size is (1n + 3)$\left|p\right|$. Here, “Exp” represents an exponential operation and “Pair” represents a pairing operation. A performance comparison of our scheme with those in references [29,30] is shown in Table 4.

In attribute-based cryptosystems, pairing and exponential operations are significantly more computationally intensive than other operations. Therefore, reducing the number of these operations can substantially improve the algorithm’s efficiency. Due to the limited use of key insulation methods, this study compares its approach with a CP-ABE method and a KP-ABE method. In our scheme, during the game encryption phase and the Site PHR encryption phase, the TA performs (2n + 2) Exp exponential operations, where n represents the number of attributes involved in the encryption. During decryption, each Site performs 1n + 1 pairing +1Exp operation. Figure 5, Figure 6, Figure 7 and Figure 8 were generated based on the calculations in Table 4. These calculations used the parameters of approximately 522 ms for an exponentiation operation and approximately 2610 ms for a pairing operation. The computations were conducted using Microsoft Excel, and the resulting graphs were created to compare the three methods.

We compare efficiency, key length, and ciphertext length from the perspectives of DRM and PHR. For DRM, our method uses a time-bound approach for encryption and computation. This method relies on hashing, resulting in low computational cost. The other two schemes do not provide this feature, making it unique to our method. For PHR, we focus on attribute-based encryption and compare the differences in efficiency, key length, and transmission volume.

We analyze the efficiency of encryption and decryption for different numbers of attributes. Figure 8 and Figure 9 show that as the number of attributes increases, our computational cost also increases. Figure 8 shows that our encryption method is more efficient than Hong et al.’s scheme but less efficient than He et al.’s scheme. The lower efficiency of our encryption method is due to the key insulation feature, which adds some computational overhead. However, this feature significantly improves private key security. In practice, the number of attributes depends on the application scenario and the number and roles of PHR administrators.

Key length affects the cost of key management. For different attributes, our private key length is equal to that of Hong et al.’s scheme but longer than He et al.’s scheme. In our method, the help key is discarded after use, ensuring private key security. Although our key length is longer than He et al.’s scheme, our method guarantees private key security and achieves long-term protection. Figure 10 shows that for four attributes, our key length is equal to Hong et al.’s scheme and shorter than He et al.’s scheme.

In Figure 11, we can see that our ciphertext length is shorter than both Hong et al.’s scheme and He et al.’s scheme. In practical applications, ciphertext size affects user experience. Larger ciphertexts result in longer waiting times. Our method ensures efficient transmission by reducing ciphertext length.

Our method provides key insulation and time-bound features compared to Hong et al.’s scheme and He et al.’s scheme. Key insulation offers better efficiency than Hong et al.’s scheme. It protects private keys and allows key updates while keeping the public key unchanged. It ensures long-term security and reduces risks from key leakage. The time-bound feature supports on-demand DRM permission management. It protects game copyrights during gameplay and allows real-time subscription and permission control. Our method offers significant advantages in security and efficiency for applications requiring DRM and PHR protection.

Overall, these figures suggest that our scheme strikes an effective balance between computational efficiency and resource requirements across encryption, decryption, and storage. Compared to other schemes, it achieves moderate overheads, supporting versatility without significant performance trade-offs.

7. Discussions

The rapid development of smart healthcare is transforming traditional medical systems into personalized, health-focused care. Customized services and gamification improve user engagement and system effectiveness. They promote healthier behaviors and reduce the burden on healthcare systems. As technology advances, smart healthcare will become increasingly important in improving public health.

Personal Health Records (PHR) allow users to access, share, and manage their health information. PHRs are essential in health promotion systems, providing personalized and convenient health management [34]. They have evolved from simple databases into platforms with features like health reminders, preventive advice, and disease monitoring [35]. These systems improve medical efficiency by reducing errors and simplifying communication with healthcare providers [36,37]. However, the widespread use of PHRs raises serious data security concerns. Risks include data breaches through hacking or social engineering [38], unauthorized access due to weak controls, and data tampering, which can result in incorrect health decisions. To address these risks, PHR systems should use strong encryption [39], multi-factor authentication [40], and role-based access controls.

Several technologies have been developed to reduce the risks of data breaches and tampering in PHRs. Encryption, both symmetric and asymmetric, ensures data confidentiality during transmission and storage, preventing unauthorized access [40]. Role-Based Access Control (RBAC) assigns permissions based on user roles, limiting access to specific data. Attribute-Based Access Control (ABAC) is more flexible, and grants access based on attributes like position or purpose, making it suitable for complex systems [14]. Blockchain technology offers a decentralized solution, ensuring data immutability and secure sharing without intermediaries. This gives users full control over their data [38]. These methods protect patient privacy, enhance data security, and allow fast access during emergencies [39,41,42].

RBAC is widely used in health promotion systems to grant access based on roles like doctors or administrators. ABAC provides finer control by granting permissions based on attributes, while encryption adds an additional layer of security, especially for cloud-based systems [14,43]. While RBAC is easier to manage, ABAC and encryption provide stronger security but require more resources for management.

Time-based access control further improves PHR security by ensuring that data are available only during specific timeframes. It prevents expired access by automatically revoking permissions when no longer needed. This method addresses dynamic demands by limiting access to relevant contexts and ensuring compliance with data retention regulations [40,44]. Time-bounded tokens and automatic revocation are key methods. Tokens grant temporary access and become invalid after expiration [17,44]. These methods are essential for preventing unauthorized access and ensuring secure data availability.

Although He et al.’s scheme [21] proposed a key escrow-free method, similar approaches cannot recover lost keys through a third party in the event of key loss, thereby increasing the burden of key management on users. Additionally, lost keys may expose users to potential attacks, compromising the overall security of the system. To address these issues, we propose a key insulation method, which mitigates the risks associated with key loss and enhances security by isolating keys and limiting the impact of compromises.

While Hong et al.’s scheme KI-CPABE-KEA [26] focuses on addressing key leakage through features like accountability, forward security, and backward security, it lacks fine-grained control over data access within specific time periods. Our scheme builds on its key insulation foundation by introducing a time-restriction mechanism, enabling precise control of user access based on designated time intervals. This feature not only ensures security in dynamic environments but also provides greater flexibility and adaptability for complex scenarios, such as those involving time-sensitive data or services.

Based on Table 5, our method offers a comprehensive solution by integrating key insulation, time-bound access control, and on-demand Data/DRM protection. This combination reduces unauthorized access risks and simplifies user management. Unlike Hong et al.’s scheme [26] and He et al.’s scheme [21], our method effectively prevents man-in-the-middle (MITM) attacks, addressing a key vulnerability in those schemes.

For replay attacks, Hong et al.’s scheme [26] does not provide protection. Both our method and He et al.’s scheme [21] ensures robust resistance. Regarding insider attacks, our method and Hong et al.’s scheme [26] offer strong safeguards, while He et al.’s scheme [21] remains vulnerable. All three methods resist collusion attacks, but He et al.’s scheme [21] lacks key leakage resistance. This feature, which is essential for long-term data security, is provided by Hong et al.’s scheme [26] and our method.

Our method ensures both perfect forward secrecy and perfect backward secrecy, matching the security levels of Hong et al.’s scheme [26] and He et al.’s scheme [21]. While all three methods include on-demand Data/DRM protection, our scheme adds time-bound authentication. This feature supports precise temporal access control, making our method particularly suitable for applications requiring time-sensitive access. These capabilities make our system more secure, versatile, and practical for real-world applications demanding high security and temporal control.

8. Conclusions

This study proposes a Key-Policy Attribute-Based encryption (KP-ABE) scheme with key insulation and time-bound encryption for a VR-based Personal Health Promotion System. KP-ABE implements fine-grained, hierarchical access control, ensuring that only authorized sites can access specific games based on their attributes. The key insulation mechanism helps prevent data theft, while the time-bound mechanism regulates the duration of authorized access for sites from different sectors or fields.

In the future, an implementation demonstration based on this algorithm can be developed, along with a performance analysis in line with the methods presented in this paper. Additionally, formal proof tools, such as Scyther tool or ProVerif, can be used to analyze the algorithm. Moreover, post-quantum cryptography could be incorporated to strengthen the security of the algorithm.

Author Contributions

C.-L.H. was responsible for method design and writing. L.-S.L. conducted security analysis and proofs. T.-L.H. and W.-Q.L. provided significant contributions to the overall writing of the paper. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding authors.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

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.

Figures and Tables

Figure 1. Key-Policy Attribute-Based Encryption.

Figure 2. Time-bound example.

Figure 3. Personal health record access based on case manager roles.

Figure 4. VR-based health promotion game subscriptions across sites.

Figure 5. System framework of a game-based Personal Health Promotion System.

Figure 6. System architecture.

Figure 7. Game demonstration.

Figure 8. Encryption computation costs [21,26].

Figure 9. Decryption computation costs [21,26].

Figure 10. Size of private key comparison [21,26].

Figure 11. Size of ciphertext comparison [21,26].

Table 1

Two types of Attribute-Based Encryption Comparison.

	KP-ABE	CP-ABE
Attribute Matching	Matches with ciphertext	Matches with private key
Access Policy Matching	Matches with private key	Matches with ciphertext
Data Owner’s Control	Cannot directly decide who obtains access	Can directly decide who obtains access
Computation Requirements	Lower	Higher
Resistance to Collusion	Lower	Higher

Table 2

SCM Access control.

Site Case Manager (SCM) Role	Prevent CVA	Health Promotion	Prevent Osteoporosis
Physical Examination Center SCM	O	O	O
CVA SCM	O	✗	✗
Health promotion SCM	✗	O	✗
Osteoporosis SCM	✗	✗	O

O indicates access is granted to the relevant data, while ✗ indicates access is denied.

Table 3

Notations of proposed scheme.

Notations	Definitions
$G_1$, $G_T$	A set of prime numbers $p$, $g$ is a generator of $G_1$
$Z_p$	The set of $\left\{0,\dots ,p-1\right\}$, $p$ is a prime number
$TK$	Time Key
SK	Secret Key
$at{t}_i$	Returns the properties linked to the leaf node
$x_s,x_e,a,v,y$	Random secret value
$u_s$	System attribute domain
td	Current time
$t{d}_1,t{d}_2$	Subscription start time and subscription end time
$t{n}_1,t{n}_2$	Subscription game start range and subscription game end range
$t_1,t_2,$…$,t_n$	Secret random values based on the attributes $u_s$
$T{K}_1,T{K}_2$	Time-bound keys
$E{K}_{u,0}$	Key insulation key
$H{K}_{u,1}$	Key insulation update help key
Pub_info, Priv_info	Public information and private information
Pub_phrinfo,Priv_phrinfo	Public content information and private content information

Table 4

Efficiency comparison.

Scheme	Encryption Costs	Decryption Costs	Size of Private Key	Size of Ciphertext
Hong et al.’s scheme [26]	(3n + 3) Exp	(3n + 2) Pair + 1nExp	(2n + 1)$\left|p\right|$	(3n + 2)$\left|p\right|$
He et al.’s scheme [21]	(1n + 4) Exp	(1n + 3) Pair + 2Exp	(1n + 5)$\left|p\right|$	(1n + 4)$\left|p\right|$
Ours	(2n + 2) Exp	(1n + 1) Pair + 1Exp	(2n + 1)$\left|p\right|$	(1n + 3)$\left|p\right|$

Table 5

Comparison of security properties.

	Hong et al.’s Scheme [26]	He et al.’s Scheme [21]	Ours
Resist man-in-the-middle (MITM) attacks	✗	✗	O
Resist replay attacks	✗	O	O
Resist insider attacks	O	✗	O
Resist collusion	O	O	O
Resist key leakage (key insulated)	O	✗	O
Provide perfect forward secrecy	O	O	O
Perfect backward secrecy	O	O	O
Provide time bound authentication	✗	✗	O
On-demand Data/DRM protection	✗	✗	O
Long-term security	O	✗	O

The symbol O denotes that the protocol achieves a specific function. We also use the symbol ✗ to denote that the function is not achieved by the protocol.

References

1. Rahmani, A.; Gia, T.N.; Negash, B.; Anzanpour, A.; Azimi, I.; Liljeberg, P. Exploiting smart e-Health gateways at the edge of healthcare Internet-of-Things: A fog computing approach. Future Gener. Comput. Syst.; 2018; 78, pp. 641-658. [DOI: https://dx.doi.org/10.1016/j.future.2017.02.014]

2. Nasr, M.M.; Islam, M.M.; Shehata, S.; Karray, F.; Quintana, Y. Smart healthcare in the age of AI: Recent advances, challenges, and future prospects. IEEE Access; 2021; 9, pp. 145248-145270. [DOI: https://dx.doi.org/10.1109/ACCESS.2021.3118960]

3. Dritsa, D.; Biloria, N. Towards a multi-scalar framework for smart healthcare. Smart Sustain. Built Environ.; 2018; 7, pp. 33-52. [DOI: https://dx.doi.org/10.1108/SASBE-10-2017-0057]

4. Capossele, A.; Gaglione, A.; Nati, M.; Conti, M.; Lazzeretti, R.; Missier, P. Leveraging blockchain to enable smart-health applications. Proceedings of the 2018 IEEE 4th International Forum on Research and Technology for Society and Industry (RTSI); Palermo, Italy, 10–13 September 2018.

5. Libanori, A.; Chen, G.; Zhao, X.; Zhou, Y.; Chen, J. Smart textiles for personalized healthcare. Nat. Electron.; 2022; 5, pp. 142-156. [DOI: https://dx.doi.org/10.1038/s41928-022-00723-z]

6. Soliman, A.H.; Ibrahim, M.; El-Hennawy, A. Improving Security and Efficiency of Enterprise Digital Rights Management. Proceedings of the 2015 6th International Conference on Computing, Communication and Networking Technologies (ICCCNT); Dallas-Fortworth, TX, USA, 13–15 July 2015; pp. 1-7.

7. Hussain, Z.; Griffiths, M.D. The Use of Digital Games for Physical Activity Promotion: A Meta-Analysis. Games Health J.; 2021; 10, pp. 225-233.

8. Johnson, D.; Deterding, S.; Grubb, M. VR Fitness Games and Their Impact on Physical Activity. J. VR Health Promot.; 2022; 8, pp. 15-27.

9. Engelberger, R.; Fetscherin, M.; Günnewig, D. Digital rights management. Wirtschaftsinformatik; 2005; 47, pp. 141-147. [DOI: https://dx.doi.org/10.1007/BF03250987]

10. Safavi-Naini, R.; Sheppard, N.P.; Uehara, T. Import/export in digital rights management. Proceedings of the 4th ACM workshop on Digital rights management (DRM ’04); Washington, DC, USA, 25 October 2004; Association for Computing Machinery: New York, NY, USA, 2004; pp. 99-110.

11. Venkatesh, B.H.; Sai, A.P.; Reddy, M.R.; Fathimabi, S. Cloud based Personal Health Record Management System and Medical Recommender System. Proceedings of the 2022 7th International Conference on Communication and Electronics Systems (ICCES); Coimbatore, India, 22–24 June 2022; pp. 1744-1749.

12. Alsyouf, A.; Lutfi, A.; Alsubahi, N.; Alhazmi, F.; Al-Mugheed, K.; Anshasi, R.J. The Use of a Technology Acceptance Model (TAM) to Predict Patients’ Usage of a Personal Health Record System: The Role of Security, Privacy, and Usability. Int. J. Environ. Res. Public Health; 2023; 20, 1347. [DOI: https://dx.doi.org/10.3390/ijerph20021347]

13. Alanazi, A.; Alanazi, M.R.; Aldosari, B. Personal Health Record (PHR) Experience and Recommendations for a Transformation in Saudi Arabia. J. Pers. Med.; 2023; 13, 1275. [DOI: https://dx.doi.org/10.3390/jpm13081275]

14. Chen, D.; Linling, C.; Xiaowei, F.; Liwen, H.; Su, P.; Ruoxiang, H. Securing patient-centric personal health records sharing system in cloud computing. China Commun.; 2014; 11, pp. 121-127. [DOI: https://dx.doi.org/10.1109/CC.2014.7022535]

15. Goyal, V.; Jain, A.; Pandey, O.; Sahai, A. Bounded Ciphertext Policy Attribute Based Encryption. Automata, Languages and Programming, Proceedings of the 35th International Colloquium, ICALP 2008, Reykjavik, Iceland, 7–11 July 2008; Springer: Berlin/Heidelberg, Germany, 2008; pp. 579-591.

16. Balabhadra, A.; Alla, R.; Mallipudi, R.K.; Bikkina, N.S.; Vurukonda, N.; Kunda, V.P. A Study On Ciphertext Policy Attribute Based Encryption. Proceedings of the 2023 9th International Conference on Advanced Computing and Communication Systems (ICACCS); Coimbatore, India, 17–18 March 2023.

17. Ragesh, G.; Baskaran, K. Cryptographically Enforced Data Access Control in Personal Health Record Systems. Procedia Technol.; 2016; 25, pp. 473-480. [DOI: https://dx.doi.org/10.1016/j.protcy.2016.08.134]

18. Li, J. Ensuring Privacy in a Personal Health Record System. Computer; 2015; 48, pp. 24-31. [DOI: https://dx.doi.org/10.1109/MC.2015.43]

19. Agrawal, S.; Chase, M. FAME: Fast Attribute-based Message Encryption. Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security; Dallas, TX, USA, 30 October–3 November 2017.

20. Ibraimi, L.; Petkovic, M.; Nikova, S.; Hartel, P.; Jonker, W. Mediated Ciphertext-Policy Attribute-Based Encryption and Its Application. Proceedings of the Information Security Applications: 10th International Workshop, WISA 2009; Busan, Republic of Korea, 25–27 August 2009.

21. He, X.; Li, L.; Peng, H. A Key Escrow-Free KP-ABE Scheme and Its Application in Standalone Authentication in IoT. IEEE Internet Things J.; 2024; 11, pp. 11381-11394. [DOI: https://dx.doi.org/10.1109/JIOT.2023.3329703]

22. Bethencourt, J.; Sahai, A.; Waters, B. Ciphertext-Policy Attribute-Based Encryption. Proceedings of the 2007 IEEE Symposium on Security and Privacy (SP ’07); Berkeley, CA, USA, 20–23 May 2007.

23. Müller, S.; Katzenbeisser, S.; Eckert, C. On multi-authority ciphertext-policy attribute-based encryption. Bull. Korean Math. Soc.; 2009; 46, pp. 803-819. [DOI: https://dx.doi.org/10.4134/BKMS.2009.46.4.803]

24. Wang, C.J.; Luo, J.F. A Key-policy Attribute-based Encryption Scheme with Constant Size Ciphertext. Proceedings of the 2012 Eighth International Conference on Computational Intelligence and Security; Guangzhou, China, 17–18 November 2012.

25. Hsu, C.-L.; Lin, H.-Y. New Identity-based Key-insulated Convertible Multi-authenticated Encryption Scheme. J. Netw. Comput. Appl.; 2011; 34, pp. 1724-1731. [DOI: https://dx.doi.org/10.1016/j.jnca.2011.06.005]

26. Hong, H.; Sun, Z.; Liu, X. A key-insulated CP-ABE with key exposure accountability for secure data sharing in the cloud. KSII Trans. Internet Inf. Syst.; 2016; 10, pp. 2394-2406.

27. Dodis, Y.; Katz, J.; Xu, S.; Yung, M. Key-insulated public key cryptosystems. Advances in Cryptology—EUROCRYPT 2002; Lecture Notes in Computer Science Springer: Berlin/Heidelberg, Germany, 2002; 2332, pp. 65-82.

28. Shikata, J.; Watanabe, Y. Identity-based encryption with hierarchical key-insulation in the standard model. Des. Codes Cryptogr.; 2018; 87, pp. 1005-1033. [DOI: https://dx.doi.org/10.1007/s10623-018-0503-4]

29. Chen, H.Y. Efficient time-bound hierarchical key assignment scheme. IEEE Trans. Knowl. Data Eng.; 2004; 16, pp. 1301-1304. [DOI: https://dx.doi.org/10.1109/TKDE.2004.59]

30. Tzeng, W.G. A Time-Bound Cryptographic Key Assignment Scheme for Access Control in a Hierarchy. IEEE Trans. Knowl. Data Eng.; 2002; 14, pp. 182-188. [DOI: https://dx.doi.org/10.1109/69.979981]

31. Chien, H.Y. Group-Oriented Range-Bound Key Agreement for Internet of Things Scenarios. IEEE Internet Things J.; 2018; 5, pp. 1890-1903. [DOI: https://dx.doi.org/10.1109/JIOT.2018.2817075]

32. Banerjee, S.; Odelu, V.; Das, A.K.; Chattopadhyay, S.; Kumar, N.; Park, Y.; Tanwar, S. Design of an Anonymity-Preserving Group Formation Based Authentication Protocol in Global Mobility Networks. IEEE Access; 2018; 6, pp. 20673-20693. [DOI: https://dx.doi.org/10.1109/ACCESS.2018.2827027]

33. Ghammam, L.; Fouotsa, E. Improving the computation of the optimal ate pairing for a high security level. J. Appl. Math. Comput.; 2019; 59, pp. 21-36. [DOI: https://dx.doi.org/10.1007/s12190-018-1167-y]

34. Tian, S.; Yang, W.; Le Grange, J.M.; Peng, W.; Ye, Z. Smart healthcare: Making medical care more intelligent. Glob. Health J.; 2019; 3, pp. 62-65. [DOI: https://dx.doi.org/10.1016/j.glohj.2019.07.001]

35. Ant Ozok, A.; Wu, H.; Garrido, M.; Pronovost, P.; Gurses, A. Usability and perceived usefulness of Personal Health Records for preventive health care. Appl. Ergon.; 2014; 45, pp. 613-628. [DOI: https://dx.doi.org/10.1016/j.apergo.2013.09.005]

36. Tang, P.; Ash, J.; Bates, D.; Overhage, J.; Sands, D.Z. Personal health records: Definitions, benefits, and strategies for overcoming barriers to adoption. J. Am. Med. Inform. Assoc.; 2006; 13, pp. 121-126. [DOI: https://dx.doi.org/10.1197/jamia.M2025]

37. Liaw, T.; Lawrence, M.; Rendell, J. The effect of a computer-generated patient-held medical record summary and/or a written personal health record on patients’ attitudes, knowledge and behaviour concerning health promotion. Fam. Pract.; 1996; 13, pp. 289-293. [DOI: https://dx.doi.org/10.1093/fampra/13.3.289]

38. Rau, H.-H.; Wu, Y.-S.; Chu, C.; Wang, F.-C.; Hsu, M.; Chang, C.-W.; Chen, K.-H.; Lee, Y.-L.; Kao, S.; Chiu, Y. et al. Importance-Performance Analysis of Personal Health Records in Taiwan: A Web-Based Survey. J. Med. Internet Res.; 2017; 19, e131. [DOI: https://dx.doi.org/10.2196/jmir.7065]

39. Señor, I.C.; Alemán, J.; Álvarez, J. Personal Health Records: New Means to Safely Handle Health Data?. Computer; 2012; 45, pp. 27-33. [DOI: https://dx.doi.org/10.1109/MC.2012.285]

40. Li, J.; Yu, Q.; Zhang, Y.; Shen, J. Key-policy attribute-based encryption against continual auxiliary input leakage. Inf. Sci.; 2019; 470, pp. 175-188. [DOI: https://dx.doi.org/10.1016/j.ins.2018.07.077]

41. Kodhai, E.; Gowri, M.; Susmitha, S.; Muthamizh, R. Literature Review on Access Control for Personal Health Records. Proceedings of the 2020 International Conference on System, Computation, Automation and Networking (ICSCAN); Pondicherry, India, 3–4 July 2020; pp. 1-5.

42. Røstad, L.; Nytrø, Ø. Personalized access control for a personally controlled health record. Proceedings of the 2nd ACM Workshop on Computer Security Architectures; Alexandria, VA, USA, 31 October 2008; pp. 9-16.

43. Sujansky, W.; Faus, S.A.; Stone, E.; Brennan, P. A method to implement fine-grained access control for personal health records through standard relational database queries. J. Biomed. Inform.; 2010; 43, (Suppl. S5), pp. S46-S50. [DOI: https://dx.doi.org/10.1016/j.jbi.2010.08.001]

44. Alhaqbani, B.; Fidge, C. Access Control Requirements for Processing Electronic Health Records. International Conference on Business Process Management; Springer: Berlin/Heidelberg, Germany, 2007; pp. 371-382.