NSG 421 NURSING INFORMATICS
FINAL EXAM
Exam Elaborations Questions &
Answers
2026
,A nurse informaticist is leading a task force to evaluate the implementation of a new
electronic health record system across a 450-bed academic medical center. Three
months post-implementation, nursing staff report that medication administration
documentation now takes an average of 22 minutes per patient per shift compared to
9 minutes with the previous system. Nurses are beginning to use workarounds
including documenting medications before administering them to save time. Using
the sociotechnical systems model as applied to health information technology, which
root cause most accurately explains this workflow breakdown and what is the most
appropriate informatics-driven response?
A. Nurses are resistant to change and need additional training on the new EHR
interface B. The EHR vendor failed to build an adequate medication administration
module and should be replaced immediately C. The implementation did not
adequately account for the interaction between the new technology and existing
clinical workflow, human factors, and organizational culture; the appropriate
response is to conduct a formal workflow analysis using time-motion study
methodology, engage frontline nurses in a redesign process using usability testing
principles, collaborate with the vendor to optimize the medication administration
workflow, and address workaround behaviors through system redesign rather than
punitive policy enforcement D. The 22-minute documentation time is acceptable
given the complexity of the new system and will normalize within six months
without intervention
Correct Answer: C
Rationale: The sociotechnical systems model, foundational to nursing informatics
theory, holds that health information technology exists within a complex web of
interactions among people, processes, technology, and organizational structures.
When technology is implemented without thorough workflow analysis and human
factors consideration, the system technically functions but creates friction that drives
workarounds. Workarounds are among the most dangerous patient safety signals in
health IT environments because they bypass safety checks including barcode
medication administration scanning, allergy alerts, and five-rights verification. The
informatics response must address the root cause at the system-workflow interface
level, not at the individual nurse compliance level. Formal usability testing and
participatory redesign with frontline nurses is the evidence-based standard.
A clinical informatics nurse specialist is reviewing alert data from her hospital's
clinical decision support system integrated within the EHR. She discovers that
2
,physicians and nurses are overriding 94% of all drug-drug interaction alerts
generated by the system over a 90-day period. The chief medical officer asks her to
explain why this pattern is dangerous even though no adverse drug events have been
documented during this period. Which explanation most accurately reflects the
informatics concept underlying this concern?
A. Alert overrides are always appropriate because clinicians are trained to use
judgment when technology generates warnings B. A 94% override rate indicates alert
fatigue, a well-documented phenomenon in which excessive, non-specific, or low-
relevance clinical decision support alerts desensitize clinicians to the point where
they habitually override all alerts including clinically critical ones without reading
them, creating a latent system failure that dramatically increases the risk of missing a
true, life-threatening drug interaction that could produce serious patient harm C. The
CDS system is malfunctioning and generating false positives that should be reported
to the vendor D. A high override rate is expected and acceptable in the first year of
CDS implementation as clinicians adjust to the new system
Correct Answer: B
Rationale: Alert fatigue is one of the most extensively documented patient safety
hazards in health information technology. Research published in journals including
JAMIA and the Journal of Patient Safety consistently demonstrates that when alert
override rates exceed 90%, clinicians are no longer reading alert content before
dismissing alerts, meaning that true, clinically significant safety warnings are being
ignored at the same rate as irrelevant ones. The absence of documented adverse
events during the observation period represents survivorship bias, not system safety.
The informatics nurse's response should include a comprehensive alert audit to
identify alerts with the highest override rates, elimination or downgrading of low-
specificity alerts, restructuring of alert presentation using principles of alert design
including tiered severity and interruptive versus non-interruptive formatting, and re-
education paired with system redesign.
A nurse informaticist at a regional health system is evaluating a proposal to
implement an artificial intelligence-based early warning system that uses machine
learning algorithms trained on three years of EHR data to predict sepsis onset up to
six hours before clinical recognition. During her evaluation she discovers that the
training data set used to build the algorithm was derived exclusively from a large
urban academic medical center with a predominantly White, insured patient
population. Her institution serves a rural, racially diverse, and largely uninsured
population. Which informatics and health equity concept most directly informs her
concern about implementing this algorithm without modification?
3
, A. Machine learning algorithms trained on EHR data are universally generalizable
because they are based on physiological parameters that do not vary by race or
insurance status B. Algorithmic bias, which occurs when a predictive model trained
on a non-representative dataset produces systematically less accurate predictions for
populations underrepresented in the training data, creating a health equity risk in
which the sepsis prediction tool may perform well for patients similar to the training
population but produce disproportionate false negatives in Black, rural, or uninsured
patients, thereby widening rather than narrowing existing disparities in sepsis
recognition and mortality C. The algorithm should be implemented immediately
because any early warning tool is better than none in a rural setting with limited
specialist access D. The concern about training data demographics is a research
consideration rather than a clinical informatics concern and should be deferred to the
research committee
Correct Answer: B
Rationale: Algorithmic bias in clinical AI is a recognized and well-documented
patient safety and health equity concern that has been formally addressed by the
FDA, the National Academy of Medicine, and the Office of the National Coordinator
for Health Information Technology. A sepsis prediction model trained exclusively on
urban, academically managed, predominantly White patients will encode the clinical
presentations, laboratory value distributions, and documentation patterns of that
population. Patients whose sepsis presentations differ systematically due to
biological, clinical, or social factors will be systematically underpredicted. The nurse
informaticist's role in AI governance includes evaluating training data
representativeness, demanding validation data from demographically comparable
populations, and advocating for equity-aware algorithm development and ongoing
surveillance of differential performance across patient subgroups.
A nursing informatics specialist is conducting a training session for newly hired
nurses on the hospital's barcode medication administration system integrated with the
EHR. During a simulation exercise, a nurse demonstrates that she obtained the
correct medication for the correct patient but bypassed the barcode scan because the
patient's wristband was damaged and the scanner produced an error. She
administered the medication correctly by memory. Using the clinical informatics
framework for patient safety technology, which response most accurately addresses
this situation?
A. Commend the nurse for clinical judgment and instruct her to replace damaged
wristbands before administering medications in the future B. Report the nurse to the
pharmacy for bypassing the barcode system C. Recognize that a workaround
occurred that bypassed a critical safety technology layer, investigate the system
4
FINAL EXAM
Exam Elaborations Questions &
Answers
2026
,A nurse informaticist is leading a task force to evaluate the implementation of a new
electronic health record system across a 450-bed academic medical center. Three
months post-implementation, nursing staff report that medication administration
documentation now takes an average of 22 minutes per patient per shift compared to
9 minutes with the previous system. Nurses are beginning to use workarounds
including documenting medications before administering them to save time. Using
the sociotechnical systems model as applied to health information technology, which
root cause most accurately explains this workflow breakdown and what is the most
appropriate informatics-driven response?
A. Nurses are resistant to change and need additional training on the new EHR
interface B. The EHR vendor failed to build an adequate medication administration
module and should be replaced immediately C. The implementation did not
adequately account for the interaction between the new technology and existing
clinical workflow, human factors, and organizational culture; the appropriate
response is to conduct a formal workflow analysis using time-motion study
methodology, engage frontline nurses in a redesign process using usability testing
principles, collaborate with the vendor to optimize the medication administration
workflow, and address workaround behaviors through system redesign rather than
punitive policy enforcement D. The 22-minute documentation time is acceptable
given the complexity of the new system and will normalize within six months
without intervention
Correct Answer: C
Rationale: The sociotechnical systems model, foundational to nursing informatics
theory, holds that health information technology exists within a complex web of
interactions among people, processes, technology, and organizational structures.
When technology is implemented without thorough workflow analysis and human
factors consideration, the system technically functions but creates friction that drives
workarounds. Workarounds are among the most dangerous patient safety signals in
health IT environments because they bypass safety checks including barcode
medication administration scanning, allergy alerts, and five-rights verification. The
informatics response must address the root cause at the system-workflow interface
level, not at the individual nurse compliance level. Formal usability testing and
participatory redesign with frontline nurses is the evidence-based standard.
A clinical informatics nurse specialist is reviewing alert data from her hospital's
clinical decision support system integrated within the EHR. She discovers that
2
,physicians and nurses are overriding 94% of all drug-drug interaction alerts
generated by the system over a 90-day period. The chief medical officer asks her to
explain why this pattern is dangerous even though no adverse drug events have been
documented during this period. Which explanation most accurately reflects the
informatics concept underlying this concern?
A. Alert overrides are always appropriate because clinicians are trained to use
judgment when technology generates warnings B. A 94% override rate indicates alert
fatigue, a well-documented phenomenon in which excessive, non-specific, or low-
relevance clinical decision support alerts desensitize clinicians to the point where
they habitually override all alerts including clinically critical ones without reading
them, creating a latent system failure that dramatically increases the risk of missing a
true, life-threatening drug interaction that could produce serious patient harm C. The
CDS system is malfunctioning and generating false positives that should be reported
to the vendor D. A high override rate is expected and acceptable in the first year of
CDS implementation as clinicians adjust to the new system
Correct Answer: B
Rationale: Alert fatigue is one of the most extensively documented patient safety
hazards in health information technology. Research published in journals including
JAMIA and the Journal of Patient Safety consistently demonstrates that when alert
override rates exceed 90%, clinicians are no longer reading alert content before
dismissing alerts, meaning that true, clinically significant safety warnings are being
ignored at the same rate as irrelevant ones. The absence of documented adverse
events during the observation period represents survivorship bias, not system safety.
The informatics nurse's response should include a comprehensive alert audit to
identify alerts with the highest override rates, elimination or downgrading of low-
specificity alerts, restructuring of alert presentation using principles of alert design
including tiered severity and interruptive versus non-interruptive formatting, and re-
education paired with system redesign.
A nurse informaticist at a regional health system is evaluating a proposal to
implement an artificial intelligence-based early warning system that uses machine
learning algorithms trained on three years of EHR data to predict sepsis onset up to
six hours before clinical recognition. During her evaluation she discovers that the
training data set used to build the algorithm was derived exclusively from a large
urban academic medical center with a predominantly White, insured patient
population. Her institution serves a rural, racially diverse, and largely uninsured
population. Which informatics and health equity concept most directly informs her
concern about implementing this algorithm without modification?
3
, A. Machine learning algorithms trained on EHR data are universally generalizable
because they are based on physiological parameters that do not vary by race or
insurance status B. Algorithmic bias, which occurs when a predictive model trained
on a non-representative dataset produces systematically less accurate predictions for
populations underrepresented in the training data, creating a health equity risk in
which the sepsis prediction tool may perform well for patients similar to the training
population but produce disproportionate false negatives in Black, rural, or uninsured
patients, thereby widening rather than narrowing existing disparities in sepsis
recognition and mortality C. The algorithm should be implemented immediately
because any early warning tool is better than none in a rural setting with limited
specialist access D. The concern about training data demographics is a research
consideration rather than a clinical informatics concern and should be deferred to the
research committee
Correct Answer: B
Rationale: Algorithmic bias in clinical AI is a recognized and well-documented
patient safety and health equity concern that has been formally addressed by the
FDA, the National Academy of Medicine, and the Office of the National Coordinator
for Health Information Technology. A sepsis prediction model trained exclusively on
urban, academically managed, predominantly White patients will encode the clinical
presentations, laboratory value distributions, and documentation patterns of that
population. Patients whose sepsis presentations differ systematically due to
biological, clinical, or social factors will be systematically underpredicted. The nurse
informaticist's role in AI governance includes evaluating training data
representativeness, demanding validation data from demographically comparable
populations, and advocating for equity-aware algorithm development and ongoing
surveillance of differential performance across patient subgroups.
A nursing informatics specialist is conducting a training session for newly hired
nurses on the hospital's barcode medication administration system integrated with the
EHR. During a simulation exercise, a nurse demonstrates that she obtained the
correct medication for the correct patient but bypassed the barcode scan because the
patient's wristband was damaged and the scanner produced an error. She
administered the medication correctly by memory. Using the clinical informatics
framework for patient safety technology, which response most accurately addresses
this situation?
A. Commend the nurse for clinical judgment and instruct her to replace damaged
wristbands before administering medications in the future B. Report the nurse to the
pharmacy for bypassing the barcode system C. Recognize that a workaround
occurred that bypassed a critical safety technology layer, investigate the system
4
