Table 2.

Taxonomy development phases.

Phase	Thematic analysis method/qualitative content analysis method	Adoption in our work
1	Reading data intensively and assessing papers	A total of 1457 papers were identified from among 2042 papers. The papers were analyzed on the basis of their titles, abstracts, and keywords, and their security, privacy, and legal features were chosen.
2	Configuring the main dimensions to correspond to the goals of this paper	The full texts of selected results in the previous phase were analyzed and processed by their details in an Excel spreadsheet. The outcome was divided into 3 categories: The first category involves the identification of privacy, security, and legal domains in the health cloud, and 150 related papers were identified. The second and third categories involve the identification of direct and indirect factors that impact privacy preservation in the health cloud. A total of 1307 remaining papers were examined by their contents.
3	Data coding in main classes	The most frequent and important features were categorized into 76 analytical categories.
4	Structuring the main classes and configuring components and subcomponents inductively on the material, and assigning subcomponents to components	The analytical categories were then synthesized into the taxonomy. The taxonomy requires a multidimensional and hierarchical structure, and each tier in the hierarchy inherits all attributes of the tier immediately above it. The highest level in the hierarchy has the greatest generality and vice versa. The subcomponents may be used to improve the domain concept under consideration and the relationships between the nodes and leaves in the hierarchy. Iterative processes can lead to taxonomy constructors. The privacy taxonomy provides a heuristic representation of hierarchies with 4 dimensions of privacy and branches in each dimension. This model allows for more specification of independent variables in the model development and with regard to the research objectives.
5	Performing category-based analyses and presenting the results	The taxonomy has 3 layers, of which the first layer has 4 main dimensions, including cloud, data, device, and legal. The second layer has 15 components, and each of them has subcomponents (n=57). This well-organized taxonomy has some advantages like presenting the hierarchical root of concepts and the inherited features of taxonomies.
6	Reporting and documentation	Finally, the taxonomy was derived and proposed from the abstraction of each of the dimensions.