UNIT 1 – HEALTHCARE DATA, INFORMATION & KNOWLEDGE
Definitions and Concepts
Health Informatics:
Health informatics is an interdisciplinary field that applies computer
science, information technology, and data analytics to improve the
communication, understanding, and management of medical and
biomedical information. It aims to enhance healthcare delivery by
leveraging real-time clinical data to support decision-making, improve
patient care, and optimize operational efficiency in healthcare systems.
Health information technology offers tools for transforming data into information that
can be utilized by clinicians and researchers. into understanding and insight. Therefore,
facilitating and enhancing human decision-making through accessible information is a
primary focus of informaticians. This concept is elaborated upon in greater detail in the
chapter addressing healthcare data, information, and knowledge.
Data, Information, and Knowledge in Healthcare:
Data refers to raw facts and figures collected from clinical
records, sensors, or other sources.
Information is data processed and organized to have meaning
and context, such as lab results linked to a specific patient.
Knowledge is synthesized information that provides insights
and understanding, enabling healthcare professionals to make
informed decisions. For example, clinical decision support
systems transform information into actionable knowledge by
correlating patient data with evidence-based guidelines
Converting Data to Information to Knowledge
Healthcare organizations collect large volumes of raw data (e.g.,
electronic health records, imaging, lab results).
This data is cleaned, integrated, and analyzed to form information that
highlights patterns or summarizations.
Advanced analytics and algorithms further synthesize information to
extract knowledge that supports diagnosis, treatment planning, and
operational improvements.
This Data-Information-Knowledge-Wisdom (DIKW) model underpins
meaningful healthcare insights and improves outcomes
Example:
Clinical Data Warehouses
A Clinical Data Warehouse (CDW) is a centralized repository that
securely stores, integrates, and manages clinical data from multiple
sources such as EHRs, laboratories, and radiology systems.
CDWs support comprehensive data analysis, clinical research, and
quality improvement by providing a unified "single source of truth" for
healthcare organizations. They enable advanced analytics,
retrospective studies, and real-time clinical decision support.
What Makes Health Informatics Difficult?
Health informatics is not just about installing technology — it’s about making
that technology work in real healthcare environments, which are complex,
unpredictable, and full of variation.
The difficulties arise from technical, organizational, and human factors.
1. Data Fragmentation
Problem: Healthcare data is scattered across different systems,
organizations, and even countries.
Why it’s difficult:
o A patient may visit multiple hospitals and clinics over time.
o Each facility may have its own EHR vendor.
o No seamless way to combine data without interoperability
standards.
Example:
A diabetic patient’s blood sugar records are in a clinic EHR, lab results
are in a hospital’s system, and medication history is in a pharmacy
database — but these don’t “talk” to each other.
2. Lack of Data Standards
Problem: Different hospitals record the same data in different formats
or codes.
Why it’s difficult:
o Hard to integrate or analyze data when units, terminologies, or
coding systems differ.
o Example: Blood glucose may be recorded in mg/dL in one lab and
mmol/L in another.
Solution attempts: Standard vocabularies like SNOMED CT, ICD-10, and
HL7 messaging formats — but adoption is uneven.
3. Data Quality Issues
Problem: Data is often incomplete, inaccurate, or outdated.
Why it’s difficult:
o Manual data entry errors.
o Missing data due to human oversight.
o Out-of-date medical histories.
Example:
A patient’s allergy information is missing from the record, leading to a
dangerous prescription error.
4. Privacy and Security Concerns
Problem: Patient data is highly sensitive and protected by strict
regulations like HIPAA (USA) and GDPR (Europe).
Why it’s difficult:
o Must ensure data confidentiality, integrity, and availability.
o Complex access controls are needed to allow the right people in
and keep the wrong people out.
Example:
A hacked hospital database can expose thousands of patient records —
leading to legal and financial consequences.
5. Interoperability Problems
Problem: Even when data exists digitally, different systems can’t
always exchange or understand it.
Why it’s difficult:
o Vendors often design systems that work only with their own
products (vendor lock-in).
o Legacy systems (old software/hardware) don’t integrate well with
modern platforms.
Example:
A hospital can’t send lab results to a referring physician’s EHR because
the file formats are incompatible.
6. Workflow Mismatch
Problem: Technology might not fit the way healthcare workers actually
work.
Why it’s difficult:
o Doctors and nurses have established workflows.
o New software may add extra steps instead of simplifying work.
Example:
A nurse must click through 10 screens to input vital signs, which takes
longer than writing them on paper.
7. Change Resistance
Problem: Healthcare staff may be reluctant to adopt new systems.
Why it’s difficult:
o Fear of change.
o Lack of training or confidence.
o Worry about increased workload.
Example:
Senior doctors who have used paper charts for 30 years may resist
using an EHR.
8. Big Data Complexity
Problem: Healthcare data is high volume, high velocity, and high
variety.
Why it’s difficult:
o Requires advanced computing power.
o Need for AI and analytics to extract useful insights.
Example:
A hospital ICU generates hundreds of data points per patient per
minute — too much for humans to analyze manually.
9. Legal and Ethical Challenges
Problem: Ethical issues arise in data use and AI decision-making.
Why it’s difficult:
o Who owns the data — patient, hospital, or software vendor?
o How to ensure AI algorithms are unbiased?
Example:
An AI that recommends treatments may unintentionally favor patients
from one demographic group over another.
Why Health IT Sometimes Fails
Poor Planning: Example – No clear workflow design before EHR
installation.
Lack of Training: Example – Staff unable to operate new software.
Technical Limitations: Example – System downtime during
surgeries.
Resistance to Change: Example – Senior doctors refusing to adopt
telemedicine.
8. Terminology of Analytics
1. Descriptive Analytics: Summarizes past data to understand trends.
Example – Hospital readmission rates last year.
2. Diagnostic Analytics: Explains why something happened. Example –
Analysis showing high infection rate due to poor sanitation.
3. Predictive Analytics: Uses historical data to predict future outcomes.
Example – Forecasting patient admission numbers.
4. Prescriptive Analytics: Recommends actions for best results.
Example – Suggesting optimal staff schedules based on patient flow.
5. Real-time Analytics: Processes data instantly. Example – Monitoring
ICU patients’ vitals live.
Recent research focuses on predictive analytics in healthcare settings to maximize health and
financial outcomes
Challenges to Data Analytics in Healthcare
1. Data Silos: Example – Patient lab reports stored in one system,
imaging data in another, with no integration.
2. Data Privacy Regulations: Example – Compliance with HIPAA or
GDPR delaying research access.
3. Unstructured Data Complexity: Example – AI struggling to analyze
handwritten doctor notes.
4. Bias in Data: Example – Predictive model underperforming for
certain ethnic groups due to underrepresentation in training data.
5. Scalability Issues: Example – Difficulty processing large-scale
national health datasets.
6. Lack of Skilled Professionals: Example – Shortage of healthcare
data scientists.
Research and Application of Analytics
The scientific foundation for using analytics to enhance healthcare delivery
remains in its nascent phase. There is a growing body of research that shows
how data from operational healthcare systems may be utilized to find people
or circumstances that are very expensive. Nonetheless, there is little
research illustrating how this data might be used to enhance healthcare
results or decrease expenses. There has been an increase in studies that use
EHR data to make clinical predictions.
Using data analytics to identify patients who are likely to be readmitted to
the hospital within 30 days after being discharged is one prominent area of
attention. The significance of this issue arises from the US Centers for
Medicare and Medicaid Services (CMS) Readmissions Reduction Program,
which penalizes hospitals for high rates of readmissions. As a result, a
number of academics have looked at EHR data to see how useful it is for
predicting which patients are likely to be readmitted.
Analytics is used extensively in improving clinical outcomes, enhancing
operational efficiency, reducing costs, and supporting public health
management.
Applications include predictive modeling for patient readmissions,
chronic disease management, resource allocation, disaster planning,
and asset tracking.
Research focuses on refining models, machine learning applications,
and developing user-friendly analytics platforms for healthcare
providers
Role of Informatics in Analytics
Informatics provides the foundational IT infrastructure, data standards,
and methodologies to capture, store, and manage healthcare data.
It enables data integration, quality control, and accessibility essential
for accurate analytics.
Informatics facilitates the translation of analytics insights into clinical
decision support and operational improvements.
By bridging clinical expertise with technology, informatics ensures
analytics tools are usable and aligned with healthcare workflows
