MIS6211 SU Data Mananagement Technical Plan for Data Governance at Yale-New Hospital Essay

1
Analysis on Data Modelling and Data Governance
Jeshion Johnson
Mis 6211 Data management
Jan 27,2022
2
Analysis on Data Modelling and Data Governance
Introduction
The company?s procedures and rules that surround data and information are commonly
referred to as “data governance,” and they are essential to the growth of any data governance
project. In addition, the business’s multiple data environments need a robust technological
infrastructure to serve them all. It is important to understand that data models may serve as a bridge
between organizational definitions and regulations and the data systems that support them. When
data governance is lacking in operational and reporting systems, these models may give useful
metadata. Before making a product available to customers, a company might utilize data modeling
to come up with a strategy. Before presenting or making a database available for usage, it is
generally agreed that the proper approach to designing relational databases is to spend time on
modeling, analysis, and a thorough understanding of the risks and problems involved. The creation
of data models and the modeling process itself has inherent value in the workplace (Katina et al.,
2020). This has been shown time and time again, particularly in the creation and delivery of
database systems, which we consider as a classic data modeling use case. Before committing to a
large-scale buildout, data modeling provides for a thorough, low-cost, and low-risk examination.
The values you establish for a company’s data governance program should be carried over.
Governance of data does not happen in a single step. A process that must be repeated over time is
required.
Oracle and Microsoft SQL Server
Customers and company alike benefit from a well-optimized database infrastructure
component, which is why database administrators work together across departments. In any
3
industry, apps and datasets are essential resources. ERP, CRM, IoT, analytics, and industryspecific (e.g., vertical) as well as DevOps are common examples of typical applications and tasks.
As a non-transactional alternative to classic data stores, NoSQL datasets are now available
alongside more standard NoSQL solutions. Business executives must keep up with the speed of
innovation and allow teams to utilize infrastructural assets, with the capital and operational aim of
lowering transaction costs.
Database are communicated with using SQL, which stands for Structured Query Language.
Database operations may be performed quickly and simply using SQL commands. The SQL query
language is known for its straightforward syntax. It is built on relational algebra and comprises
mostly of English language parts. SQL is a standardized language that may be used on a wide
variety of platforms and databases. There are several SQL-enabled databases out there (Leonelli,
2019). The common backend language may be used to access the data contained in these systems.
It is easy to execute SQL statements since just a few distinct instructions are required for most of
the statements. SQL provides features for building data structures and accessing databases in
conjunction to procedures for generating, modifying, and deleting data.
SQL programmers are largely responsible for establishing, developing, and managing SQL
databases, as well as assuring their stability and performance. For the most part, they are in charge
of putting together the application’s tables, architectures, algorithms, and dictionary. As a result of
their work, database developers contribute to the authenticity and accuracy of the system they work
with. Data security and software management validation are two of the primary responsibilities of
a SQL Server Developer. Algorithms, scripts, procedures, prompts, and procedures that enable
software development are really a part of the job description. An SQL programmer also does
database testing, problem patches, performance fixes, and solutions for the implementation of
4
corrective measures. Database permission and memory management are also part of their remit,
as are routine tasks like configuration database, restore, and update management.
However, the ability to design and change databases and write complicated SQL queries
are both prerequisites for working as a SQL developer. It is also a benefit if employees have
worked with an IDE like Oracle SQL Developer, MySQL, and MS Windows Server (Sjoukema et
al., 2021). However, it is worth noting that focusing on a single database environment is frequently
preferable than having a general understanding of numerous databases.
Best Practices About Data Modelling and Governance
Data was traditionally saved in large log files known as shortcut files, which were used to
store records in a flat database design. Many different bits of information might be found in a
single file, whether it was about people, resources, or things around the office. The data was
organized into a record, even though it was inconsistent. A lack of organization made it difficult
and time-consuming for companies to find particular records and build customized reports. This
implies that the data modelling tool that should be used has to be easy to install and use. In addition
to being simple to use, Microsoft SQL may be setup through a wizard. Unlike some other data
5
stores, SQL server has a browser configuration interface. In addition to the simple one-click
installation, it has a user-friendly interface and a wealth of documentation. Mandatory updates are
computed directly via the installation wizard, which decreases the amount of human effort required
to complete the installation. As a result of patch management, system administration costs are
reduced and the database is kept up to date with current trends. Storage and analytical services
might be set up independently at a later time.
Notably, it needs to have enhanced performance. Data processing and encrypting are builtin capabilities in SQL server, which improves speed. Applications need not be modified to protect
and encrypt data. Authorization management features and access restrictions meant to protect
critical corporate information are provided by SQL Server. MS SQL Server is available in a variety
of versions to meet the demands of both large corporations and small businesses (Lovell, Watson,
& Hiteva, 2022). The pricing and features of each version are different. As a result, businesses
have the option to choose the version that best suits their operating requirements.
For instance, the enterprise version would be most applicable in this case study. This
version is often used by bigger organizations with more storage needs. Web-based databases and
data storage are two of its key features. The fundamental characteristics of a corporate SQL server
may be found in this kind of server. Advanced encryption methods make it almost hard to breach
the security of the SQL Server database. Extra security measures have been added to SQL Server
to limit the danger of assaults. It is important to note that SQL Server has a number of advanced
tools that may be used to restore and recover data that has been lost or destroyed. The whole
database may be recovered using specialized recovery techniques. Many large enterprises rely on
SQL Server’s built-in capabilities.
6
Technical Plan for Data Architecture Management
Data architecture encompasses platform options including on computer servers, statistical
models, cloud technology, and related data operations and network setups. Databases, cloud
hosting systems, and textual and comma-separated data files are only some of the backup and
recovery systems and databases that are covered in this area of study. Effectively communicating
the advantages of establishing data architecture to senior executives is just as important as any
other strategic technology investment. Make a case for why a data architecture makes sense for
your company (Katina et al., 2020). Obtain the support of important stakeholders by identifying
and including them in the process.
The architecture will involve identification of the key personas in the organization. The
technological environment of an institution is influenced by the information demands of data
consumers. Applications are held responsible for the data they create and utilize. Determine who
in the company is responsible for creating, storing, updating, reading, and otherwise handling data.
Identify and classify archetypal personalities based on their data touchpoints. Afterwards, it will
be important to determine the information requirements. This will involve working with those who
will be using the data to learn about their business goals and how they plan to use it. Furthermore,
it will be good to write out how these needs connect to different types of data domains and distinct
datasets that these customers are now using or may need in the future.
Conclusion
In conclusion, data modelling and governance is critical for every organization.
Organizations may keep tabs on different databases, including Oracle, MySQL, SQL Server, and
cloud servers, with a single monitoring system. Among its many impressive features are
7
customized warnings, the ability to quickly spot performance problems, 24-hour monitoring, and
an intuitive user interface. Using a cloud – based architecture, DPA not only improves database
performance on VMware, AWS, and Azure, but it also speeds up development and scales more
easily. Additionally, it evaluates many variables such as activities, SQL operations, memory use,
and wait kinds together with historical data to uncover the underlying cause of a problem, as well
as to detect and avoid abnormalities in the future. As a further benefit, the response time analysis
may assist identify slowdowns in applications and databases. A collecting data engine provides
organizations with a complete and graphical overview of the data it collects.
8
References
Katina, P. F., Tolk, A., Keating, C. B., & Joiner, K. F. (2020). Modelling and simulation in
complex
system
governance. International
Journal
of
System
of
Systems
Engineering, 10(3), 262-292.
Leonelli, S. (2019). Data governance is key to interpretation: Reconceptualizing data in data
science. Harvard Data Science Review, 1(1).
Lovell, K., Watson, J., & Hiteva, R. (2022). Infrastructure decision-making: Opening up
governance futures within techno-economic modelling. Technological Forecasting and
Social Change, 174, 121208.
Sjoukema, J. W., Samia, J., Bregt, A. K., & Crompvoets, J. (2021). Governance interactions of
spatial data infrastructures: an agent-based modelling approach. International Journal of
Digital Earth, 14(6), 696-713.
1
Data Management in Clinical Effectiveness
Jeshion Johnson
South University
Data Management su01
Instructor’s Name and Title
Jan 20,2022
2
Yale-New Haven Hospital Data Management in Clinical Effectiveness
Robust information infrastructure that informs healthcare effectiveness, in reality, requires
the adoption of electronic health records and an increase in emphasizing the integration and reuse
of clinical care and administration. This infrastructure ensures that data can be stored and retrieved
to help create analytic datasets that understand the anticipated data. For Yale-New Haven Hospital
to develop such kind of infrastructure, detailed cases described the criteria, cohort selection,
research hypotheses, and the analytical plans that required the support of the database required the
active engagement of clinical investigators (Kent et al., 2021). Among the needs that claim
Comparative Effective Research investigators focused on, among other considerations, involves
the suitability of data models. In line with the hospital’s mission of its commitment to innovation
and excellence in patient care, teaching, research, and services to communities, data management
requires several processes consideration.
The processes include determining which data elements need to be stored and how they
will be stored. Such elements to be stored include their constraints and relationships. Also, to
model data management, compromises between usability and complexity which are crucial to
modelling decisions, need to be addressed. Thus, the case study for this report uses a cohort of
subjects to illustrate the need to identify patients by diagnoses, medication and age use while
excluding those with other diagnoses that may be misdiagnosed.
Need for Data Model
Several data models need to be explored against a set of investigative requirements and
technical requirements to select a data model that best fits the need of the hospital and that is
flexible to expand upon new requirements. Other valid options exist before settling on a particular
3
model, and the prioritization of requirements must depend on many factors. Other needs include
the rapid rise of the adoption of electronic health record systems, spurred by the financial
incentives and penalties established in the American Recovery and Reinvestment act. Also, there
has been a growth in interest in evaluating the effectiveness of clinical care practices that use
electronic data recorded in routine clinical care. Hence the need for a data management model is
necessitated by the keen interest in clinical effectiveness, benefits and risks of medical care in the
world of practice using administrative, clinical and billing data.
The Healthcare Quality and Research agency funded multiple projects to the tune of $1.1
billion centred on building scalable distributed research networks to promote CER as part of the
American Recovery and Reinvestment Act (De et al., 2021). SAFTINet was among the many
beneficiaries of the funding. For SAFTINet to support CER, its goal was to build a well-distributed
network focusing on safety-net stakeholders, including persons with Medicaid, persons lacking
health insurance and State Children’s Health Insurance Programs (Kent et al., 2021). The model
was supposed to combine detailed financial and clinical data from state Medicaids, health records,
and administrative data sources into a secondary analytic-only database different from clinical
application databases. And also, the model was to answer queries regarding the comparative
effectiveness of diagnostics, protocols, treatments and other delivery system factors. The hospital
needed to adopt or develop a data model for storing, processing, analyzing, and evaluating a wide
range of financial and clinical data. SAFTInet, as a case study, describes the considerations that
can be applied during the hospital’s quest to identify and develop compatible, consistent, easy to
use, scalable, secure, testable, and its functional features support the hospital activities its mission.
4
Features and Relationships
Data models are developed considering their features and relationships to the entity and its
operations. The features are data types, including date, integer, character and time. The second
feature is the constraints which look into which values are allowed and if each of the values must
be unique. Thirdly the relationship between raws of data, for example, if a row in one table can be
related to another or many rows in another table and finally if hierarchies that define concepts can
be presented (El Emam et al., 2020). Lastly, the model looks into metadata definitions,
assumptions, and procedures that explain each data element’s intended meaning and use, data
collection, allowed ranges or values, and interdependence between the data elements.
The research data that can be stored, how the values of data should be interpreted and how
easily data subsets should be extracted and queried from the research database is greatly influenced
by the structure and metadata definitions in a data model. Regardless of computer and information
sciences-oriented literature on data modelling, there are no published or well-studied options,
choices, or impact of modelling decisions that support clinical research.
Data Model Development
A data model’s structural components are typically conveyed schematically in drawings
that use symbols and notations to denote the features and relationships among data items. The
model figure should show the format of the Entity-Relationship Diagram. The model shows each
item and its data type, including character, date and integer. It also notes if the data elements are
always required to have a value “NN= never null/missing and if it is a primary or foreign key. For
example, if a patient table in a model can have a data item named “PAT_ID”, that is an integer,
but it must always have a value NN, and it should be a PK (El Emam et al., 2020). The Entity-
5
Relationship Diagram captures the database structure but not the metadata often captured in related
documents linked to an ERD.
Although the integers are simple, the ERDs usually illustrate different assumptions in
organized data and related items in a data model. Developing the model should encompass all
diagnoses, procedures, and medication prescriptions associated with the visit, especially when the
model is visiting centric. Such a model doesn’t contain a data field obtained from the required
associated visit. The model will show that the visit does not always need an associated prescription
or procedure when basing the symbols at the end of the connected lines,i.e., circles, lines, and
crow’s feet. Still, it must have at least have one ICD-9_CM diagnosis (De et al., 2021). A patientcentric model associates visits, prescriptions, and procedures to a patient instead of a visit. In such
a model, procedures and prescriptions contain fields of the date that are independent. The data in
this field can be unrelated to any patient visit, thus allowing for prescriptions and procedures to be
recorded in instances where there is no corresponding list. There is only one single date for a visit
filled in the previously described model, whereas, in the second, there are two date fields for visits.
The model should allow for hospital admission, discharge, and the decision to store both visit
dates. According to Lai et al. (2018) diagnosis codes such as ICD-9-CM, ICD-10-CM and
SNOMEd should be implemented to enable different values in Code_Type to differentiate coding
systems
ERDs are used to evaluate a proposed data model’s ability to realize the data storage and
query needs of a research project used by multidisciplinary teams, including clinical investigators,
informatics professionals, and biostatisticians. A crucial challenge for CER in putting up its
projects is designing a database structure. The database structure should be flexible to promote
various data types collected from electronic health records, pharmacy dispensing, billing systems,
6
pharmacy claims/benefits processing systems. All of these involves a process called data
integration (Choi et al., 2020). Data integration involves combining related data from separate
sources with different variable formats, data structures, specificity, quality and definitions. To
integrate data, careful special attention to how data from various systems will be presented in the
research data model and how changes in procedures and data definitions and sources will be
resolved to secure compatibility of the resulting data values. Because data model structures and
integration decisions and definitions are closely intertwined, the decisions are well addressed when
investigators and analytics work closely with database designers. They work together in exploring
the impacts of decisions on database maintenance and extensibility, data quality, accessibility and
analytic capability.
The data model should be able to embody different assumptions. The assumptions include
all actions that matter occurs during the visit. The model should be more general in other
circumstances as it represents actions between and during a visit (El Emam et al., 2020). The model
should also perform comparisons of dates between visits and prescriptions to find prescriptions
recorded on the date visit. Though balancing flexibility and complexity in data modelling is
challenging, the model data should incorporate both flexibility and complexity.
7
Figure 1: Data Model Diagram
Data model Evaluation
A comprehensive framework to ensure all potential data model features are considered in
developing a data model. Eight dimensions are proposed to be considered in evaluating a data
model. The Moody and Shanks framework emphasizes integrated analysis in which each setting
considered some factors to be crucial. the dimensions include completeness, integrity and
understandability
The model SAFTINet model should also support the following:
? Extraction of patient-level data and create analytic datasets
? Calculate prescribed drug intervals, often known as drug exposures
? Calculate ages to smaller units of measurement for young children and the year for adults
depending on age.
? Be able to link patient data across separate data sources.
8
? Identification of patients as part of a defined cohort and give way for descriptive data
collection identify a patient as part of a defined cohort to allow prospective data collection.
And support identified data with HIPAA regulations.
In conclusion, the data model application should provide access to authorized users. The
platform should be flexible to allow users to add new functions. Hence the application and platform
incorporate software components that can interact with the data model
9
References
Choi, S. A., Kim, H., Kim, S., Yoo, S., Yi, S., Jeon, Y., … & Kim, K. J. (2020). Analysis of
antiseizure drug-related adverse reactions from the electronic health record using the
common data model. Epilepsia, 61(4), 610-616.
De, A., Huang, M., Feng, T., Yue, X., & Yao, L. (2021). Analyzing Patient Secure Messages Using
a Fast Health Care Interoperability Resources (FIHR)?Based Data Model: Development
and Topic Modeling Study. Journal of medical Internet research, 23(7), e26770.
El Emam, K., Mosquera, L., & Bass, J. (2020). Evaluating identity disclosure risk in fully synthetic
health data: model development and validation. Journal of medical Internet
research, 22(11), e23139.
Kent, S., Burn, E., Dawoud, D., Jonsson, P., ?stby, J. T., Hughes, N., … & Bouvy, J. C. (2021).
Common Problems, Common Data Model Solutions: Evidence Generation for Health
Technology Assessment. PharmacoEconomics, 39(3), 275-285.
Lai, E. C. C., Ryan, P., Zhang, Y., Schuemie, M., Hardy, N. C., Kamijima, Y., … & Setoguchi, S.
(2018). Applying a common data model to Asian databases for multinational
pharmacoepidemiologic studies: opportunities and challenges. Clinical epidemiology, 10,
875.
1
Data Management in Clinical Effectiveness
South University
Data Management su01
Instructor’s Name and Title
Jan 20,2022
2
Yale-New Haven Hospital Data Management in Clinical Effectiveness
Robust information infrastructure that informs healthcare effectiveness, in reality, requires
the adoption of electronic health records and an increase in emphasizing the integration and reuse
of clinical care and administration. This infrastructure ensures that data can be stored and retrieved
to help create analytic datasets that understand the anticipated data. For Yale-New Haven Hospital
to develop such kind of infrastructure, detailed cases described the criteria, cohort selection,
research hypotheses, and the analytical plans that required the support of the database required the
active engagement of clinical investigators (Kent et al., 2021). Among the needs that claim
Comparative Effective Research investigators focused on, among other considerations, involves
the suitability of data models. In line with the hospital’s mission of its commitment to innovation
and excellence in patient care, teaching, research, and services to communities, data management
requires several processes consideration.
The processes include determining which data elements need to be stored and how they
will be stored. Such elements to be stored include their constraints and relationships. Also, to
model data management, compromises between usability and complexity which are crucial to
modelling decisions, need to be addressed. Thus, the case study for this report uses a cohort of
subjects to illustrate the need to identify patients by diagnoses, medication and age use while
excluding those with other diagnoses that may be misdiagnosed.
Need for Data Model
Several data models need to be explored against a set of investigative requirements and
technical requirements to select a data model that best fits the need of the hospital and that is
flexible to expand upon new requirements. Other valid options exist before settling on a particular
3
model, and the prioritization of requirements must depend on many factors. Other needs include
the rapid rise of the adoption of electronic health record systems, spurred by the financial
incentives and penalties established in the American Recovery and Reinvestment act. Also, there
has been a growth in interest in evaluating the effectiveness of clinical care practices that use
electronic data recorded in routine clinical care. Hence the need for a data management model is
necessitated by the keen interest in clinical effectiveness, benefits and risks of medical care in the
world of practice using administrative, clinical and billing data.
The Healthcare Quality and Research agency funded multiple projects to the tune of $1.1
billion centred on building scalable distributed research networks to promote CER as part of the
American Recovery and Reinvestment Act (De et al., 2021). SAFTINet was among the many
beneficiaries of the funding. For SAFTINet to support CER, its goal was to build a well-distributed
network focusing on safety-net stakeholders, including persons with Medicaid, persons lacking
health insurance and State Children’s Health Insurance Programs (Kent et al., 2021). The model
was supposed to combine detailed financial and clinical data from state Medicaids, health records,
and administrative data sources into a secondary analytic-only database different from clinical
application databases. And also, the model was to answer queries regarding the comparative
effectiveness of diagnostics, protocols, treatments and other delivery system factors. The hospital
needed to adopt or develop a data model for storing, processing, analyzing, and evaluating a wide
range of financial and clinical data. SAFTInet, as a case study, describes the considerations that
can be applied during the hospital’s quest to identify and develop compatible, consistent, easy to
use, scalable, secure, testable, and its functional features support the hospital activities its mission.
4
Features and Relationships
Data models are developed considering their features and relationships to the entity and its
operations. The features are data types, including date, integer, character and time. The second
feature is the constraints which look into which values are allowed and if each of the values must
be unique. Thirdly the relationship between raws of data, for example, if a row in one table can be
related to another or many rows in another table and finally if hierarchies that define concepts can
be presented (El Emam et al., 2020). Lastly, the model looks into metadata definitions,
assumptions, and procedures that explain each data element’s intended meaning and use, data
collection, allowed ranges or values, and interdependence between the data elements.
The research data that can be stored, how the values of data should be interpreted and how
easily data subsets should be extracted and queried from the research database is greatly influenced
by the structure and metadata definitions in a data model. Regardless of computer and information
sciences-oriented literature on data modelling, there are no published or well-studied options,
choices, or impact of modelling decisions that support clinical research.
Data Model Development
A data model’s structural components are typically conveyed schematically in drawings
that use symbols and notations to denote the features and relationships among data items. The
model figure should show the format of the Entity-Relationship Diagram. The model shows each
item and its data type, including character, date and integer. It also notes if the data elements are
always required to have a value “NN= never null/missing and if it is a primary or foreign key. For
example, if a patient table in a model can have a data item named “PAT_ID”, that is an integer,
but it must always have a value NN, and it should be a PK (El Emam et al., 2020). The Entity-
5
Relationship Diagram captures the database structure but not the metadata often captured in related
documents linked to an ERD.
Although the integers are simple, the ERDs usually illustrate different assumptions in
organized data and related items in a data model. Developing the model should encompass all
diagnoses, procedures, and medication prescriptions associated with the visit, especially when the
model is visiting centric. Such a model doesn’t contain a data field obtained from the required
associated visit. The model will show that the visit does not always need an associated prescription
or procedure when basing the symbols at the end of the connected lines,i.e., circles, lines, and
crow’s feet. Still, it must have at least have one ICD-9_CM diagnosis (De et al., 2021). A patientcentric model associates visits, prescriptions, and procedures to a patient instead of a visit. In such
a model, procedures and prescriptions contain fields of the date that are independent. The data in
this field can be unrelated to any patient visit, thus allowing for prescriptions and procedures to be
recorded in instances where there is no corresponding list. There is only one single date for a visit
filled in the previously described model, whereas, in the second, there are two date fields for visits.
The model should allow for hospital admission, discharge, and the decision to store both visit
dates. According to Lai et al. (2018) diagnosis codes such as ICD-9-CM, ICD-10-CM and
SNOMEd should be implemented to enable different values in Code_Type to differentiate coding
systems
ERDs are used to evaluate a proposed data model’s ability to realize the data storage and
query needs of a research project used by multidisciplinary teams, including clinical investigators,
informatics professionals, and biostatisticians. A crucial challenge for CER in putting up its
projects is designing a database structure. The database structure should be flexible to promote
various data types collected from electronic health records, pharmacy dispensing, billing systems,
6
pharmacy claims/benefits processing systems. All of these involves a process called data
integration (Choi et al., 2020). Data integration involves combining related data from separate
sources with different variable formats, data structures, specificity, quality and definitions. To
integrate data, careful special attention to how data from various systems will be presented in the
research data model and how changes in procedures and data definitions and sources will be
resolved to secure compatibility of the resulting data values. Because data model structures and
integration decisions and definitions are closely intertwined, the decisions are well addressed when
investigators and analytics work closely with database designers. They work together in exploring
the impacts of decisions on database maintenance and extensibility, data quality, accessibility and
analytic capability.
The data model should be able to embody different assumptions. The assumptions include
all actions that matter occurs during the visit. The model should be more general in other
circumstances as it represents actions between and during a visit (El Emam et al., 2020). The model
should also perform comparisons of dates between visits and prescriptions to find prescriptions
recorded on the date visit. Though balancing flexibility and complexity in data modelling is
challenging, the model data should incorporate both flexibility and complexity.
7
Figure 1: Data Model Diagram
Data model Evaluation
A comprehen