Health-System Processes and Information Requirements

Introduction

A hospital is not merely a collection of departments, technologies, and professionals. It is a network of interdependent processes through which people seek care, are assessed, diagnosed, treated, monitored, discharged, and followed up. Each process generates information, consumes information created elsewhere, and produces information needed by other participants. Hospital Information Systems (HIS) therefore succeed or fail partly according to how accurately they represent real work.

A patient journey may cross registration, triage, consultation, laboratory and imaging services, medication management, nursing care, billing, discharge planning, and community follow-up. The journey is usually not linear. It includes queues, repeated assessments, handovers, exceptions, referrals, and parallel activities. A process-oriented view enables managers and system designers to identify where information is created, who requires it, when it is required, in what form, and with what degree of reliability.

This chapter develops a practical and analytical framework for mapping health-system processes and translating them into information requirements. It treats workflow redesign as a socio-technical task: technology matters, but so do professional judgment, organizational policy, regulation, incentives, workload, and patient experience.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Map an end-to-end patient journey and distinguish major clinical, operational, and administrative processes.
2. Explain the information requirements of clinical, nursing, diagnostic, pharmacy, administrative, and managerial workflows.
3. Identify bottlenecks, dependencies, handoffs, control points, and failure modes.
4. Use process-mapping and requirements-analysis techniques to support HIS design and redesign.
5. Evaluate workflow changes using safety, quality, efficiency, user, equity, and governance criteria.
6. Apply these concepts to an integrated hospital case.

1. Health-System Processes as Socio-Technical Systems

A process is a structured set of activities that transforms inputs into outputs for a patient, professional, department, or organization. In health care, inputs include patient information, clinical observations, specimens, orders, equipment, staff time, medicines, and financial authorization. Outputs include diagnoses, treatment decisions, results, referrals, prescriptions, bills, reports, and improved health outcomes.

A useful distinction is between:

• Patient-facing processes: access, registration, triage, consultation, treatment, discharge, and follow-up.
• Clinical support processes: laboratory, imaging, pharmacy, sterile services, blood bank, and rehabilitation.
• Administrative processes: scheduling, eligibility verification, authorization, billing, coding, claims, and records management.
• Management processes: capacity planning, quality improvement, risk management, workforce planning, budgeting, and strategic decision-making.
• Enabling processes: information technology, procurement, facilities, infection prevention, maintenance, and governance.

These categories overlap. For example, a diagnostic result is a clinical output, a laboratory workload measure, a billing input, and a managerial performance indicator.

1.1 Process, workflow, pathway, and journey

A workflow describes the sequence and responsibility for activities. A care pathway describes an agreed pattern of care for a condition or patient group. A patient journey describes the patient’s experience across the whole system, including non-clinical encounters and transitions. A process map is a representation of activities, decisions, actors, information, and flows.

The same patient journey can be represented at different levels. A high-level map may show access → assessment → diagnosis → treatment → discharge. A detailed map may show identity verification, allergy confirmation, specimen labeling, result validation, clinician review, patient notification, and follow-up scheduling. The appropriate level depends on the decision being supported. Excessive detail can obscure the main problem; insufficient detail can conceal safety-critical dependencies.

1.2 The information perspective

For every process, six questions should be asked:

1. What information is needed?
2. Who needs it?
3. At what point in the process is it needed?
4. Where is it created and where is it stored?
5. What action or decision does it support?
6. What happens if it is missing, late, inaccurate, or inaccessible?

Information requirements include more than data fields. They include timing, format, usability, provenance, access rights, completeness, accuracy, and retention. A medication order, for example, requires patient identity, medicine, dose, route, frequency, indication, prescriber identity, time, and status. It also requires decision support, allergy information, interaction checking, and a mechanism for pharmacy and nursing staff to see the authorized order.

2. Mapping the Patient Journey

2.1 A generic longitudinal journey

A generalized hospital journey consists of the following stages:

1. Access and referral
2. Registration and identity management
3. Triage and initial assessment
4. Clinical consultation and history-taking
5. Diagnostic investigation
6. Diagnosis and treatment planning
7. Treatment, monitoring, and reassessment
8. Admission or continuation as an outpatient
9. Discharge, transfer, or referral
10. Follow-up and information exchange with other providers

The sequence is not necessarily linear. Diagnostic testing may lead to a revised history, an urgent escalation may bypass routine scheduling, and a patient may move from outpatient care to emergency admission. A robust HIS must support both the expected pathway and clinically justified deviations.

Figure: A patient journey supported by a shared information layer and cross-functional workflows.

2.2 Swim-lane mapping

A swim-lane map allocates activities to actors or organizational units. Typical lanes include patient, registration clerk, nurse, physician, laboratory, radiology, pharmacy, finance, medical records, and external provider. The map should show:

• activity sequence;
• decision points and exception routes;
• inputs and outputs;
• systems used;
• handoffs and waiting states;
• responsibility and authorization;
• information created, modified, or consumed.

A good map distinguishes value-adding work from necessary control work and avoidable waste. Identity verification may not directly improve a symptom, but it is necessary for safety and legal accountability. Re-entering the same demographic data into three systems, by contrast, may be avoidable waste.

2.3 Patient journey versus institutional process

The patient experiences the entire journey, whereas departments often optimize only their own segment. A laboratory may measure turnaround time from specimen receipt to result validation, while the patient waits from test ordering to clinical action. Both measures are useful, but they answer different questions. HIS analysis must therefore join departmental measures into an end-to-end view.

3. Clinical and Nursing Workflows

3.1 Clinical workflow

Clinical workflow commonly includes information gathering, clinical reasoning, ordering, interpretation, treatment selection, communication, documentation, and review. Core information requirements include:

• reliable patient identity and encounter context;
• presenting complaint, history, examination, allergies, and risk factors;
• prior diagnoses, procedures, medications, and results;
• structured and narrative documentation;
• orders with indications, priority, timing, and author;
• results with reference information, status, units, and verification;
• decision rationale, consent, care plan, and follow-up;
• alerts that are relevant, timely, and actionable.

Clinical systems should support professional judgment rather than turn every decision into a rigid sequence. Structured data improves retrieval and analysis, but narrative text remains important for uncertainty, context, and reasoning. A balanced design combines structured elements, controlled vocabularies, and appropriate free text.

3.2 Nursing workflow

Nursing work is continuous, distributed, and highly dependent on observation and coordination. It includes admission assessment, risk screening, care planning, medication administration, monitoring, patient education, escalation, shift handover, and discharge preparation.

Nursing information requirements include vital signs, pain, mobility, nutrition, skin integrity, falls risk, pressure injury risk, continence, psychosocial needs, devices, fluid balance, nursing interventions, response to treatment, and escalation events. Information must be available at the bedside and during handover. Excessive documentation burden can reduce time for direct care and encourage copying, delayed entry, or workarounds.

A nursing record should make the current state and required next actions visible. For example, a warning that a patient is at high risk of falling is useful only if it is linked to an observable care plan, responsible staff, reassessment interval, and escalation process.

3.3 Handoffs and dependencies

Handoffs are high-risk information transitions. A handoff should communicate the patient’s identity, current condition, relevant history, clinical concerns, pending actions, contingency plans, and responsibility. The system should support a shared view, but electronic visibility does not eliminate the need for professional communication. A result may be present in the record yet not recognized by the clinician responsible for acting on it.

4. Diagnostic Workflows

Diagnostic services transform a clinical question into an interpreted result. A generic diagnostic workflow includes:

1. order entry and clinical indication;
2. scheduling or prioritization;
3. patient preparation and consent where required;
4. specimen collection or image acquisition;
5. labeling and accessioning;
6. analysis or interpretation;
7. technical and clinical validation;
8. result release;
9. clinician review and patient communication;
10. follow-up or treatment action.

The workflow has important dependencies. The laboratory depends on correct patient identification and specimen labeling. Radiology depends on appropriate protocol selection, patient preparation, equipment availability, and reporting capacity. Clinicians depend on result visibility and notification of critical findings.

Diagnostic information requirements include order priority, specimen type, collection time, collector identity, chain of custody where applicable, instrument or modality, result status, units, reference intervals, abnormal flags, critical-result escalation, report author, and amendment history. The system must distinguish preliminary, corrected, and final results. Overwriting an original result without an audit trail creates clinical and legal risk.

Diagnostic bottlenecks often arise from mismatch between demand and capacity, batching, incomplete orders, transport delays, repeat collection, equipment downtime, and report backlogs. A process map should identify whether the delay occurs before collection, during analysis, during validation, or after result release.

5. Pharmacy and Medication-Use Processes

Medication management is a longitudinal process rather than a single prescribing event. It includes medication reconciliation, prescribing, verification, dispensing, administration, monitoring, discharge prescribing, and post-discharge review.

Information needed includes medicine identity, strength, dose, route, frequency, duration, indication, start and stop dates, allergies, renal or hepatic considerations, interactions, formulary status, substitution rules, prescriber authority, administration record, and patient understanding. Medication reconciliation compares the medicines a patient was taking before admission with current orders and the discharge regimen. It is particularly important at transitions, where discrepancies are common.

Computerized provider order entry and pharmacy systems can reduce illegibility and support checking, but they can also introduce new risks. Poorly designed alerts create alert fatigue; confusing order sets can produce selection errors; default values may be accepted without adequate review. Decision support should be risk-sensitive, transparent, and designed around the actual medication-use workflow.

The medication administration record should connect the authorized order to the patient, medicine, dose, route, time, administering professional, and outcome. A system should make it possible to distinguish an administered dose, a withheld dose, a refused dose, a delayed dose, and a documentation omission.

6. Administrative and Managerial Workflows

6.1 Administrative workflows

Administrative processes establish the institutional context in which clinical care occurs. Important workflows include appointment management, registration, identity creation, eligibility verification, referral authorization, billing, coding, claims submission, discharge disposition, and records release.

Administrative data must be accurate because errors propagate. A duplicate patient record can fragment allergies, results, and history. An incorrect encounter type can affect billing, statistics, and service reporting. A missing authorization may delay care even when clinical capacity exists.

The principle of single source of truth is useful but must be applied carefully. A patient’s legal name, preferred name, address, insurance status, and clinical diagnosis may have different authoritative sources and update rules. Governance should specify ownership, synchronization, correction, and conflict resolution.

6.2 Managerial workflows

Managers require information for operational control, resource allocation, quality improvement, and strategic planning. Examples include:

• bed occupancy, admissions, discharges, and transfers;
• emergency department arrivals and waiting times;
• operating-room utilization and cancellation rates;
• diagnostic turnaround time;
• medicine stock and expiry risk;
• staffing, workload, overtime, and skill mix;
• infection, incident, readmission, mortality, and patient-experience measures;
• revenue, cost, denial, and claims indicators.

A managerial dashboard should not simply display every available metric. Each measure needs a defined numerator, denominator, population, time period, data owner, refresh frequency, and interpretation. A rate without a denominator can be misleading. For example, the number of medication incidents should be interpreted alongside doses administered or patient-days, while recognizing that under-reporting can distort comparisons.

7. Requirements Analysis for Hospital Information Systems

7.1 Functional and non-functional requirements

A functional requirement states what the system must do: register a patient, transmit a laboratory order, issue a critical-result alert, or generate a discharge summary. A non-functional requirement states how the system must perform: availability, response time, usability, interoperability, security, accessibility, auditability, and maintainability.

Requirements should be written in operational terms. “The system should be user-friendly” is weak. “A nurse should be able to document a routine observation at the bedside using no more than a defined number of navigation steps, with offline recovery if connectivity is interrupted” is more testable.

7.2 Information-quality dimensions

Relevant dimensions include:

• accuracy: correspondence with reality;
• completeness: required elements are present;
• timeliness: available when needed;
• consistency: compatible across records and systems;
• validity: conforms to defined formats and values;
• provenance: origin and transformation are known;
• accessibility: authorized users can retrieve it;
• confidentiality: unauthorized disclosure is prevented.

Information quality is contextual. A result delivered two hours late may be acceptable for routine monitoring but unsafe for a critical emergency test. A free-text note may be rich in context but difficult to aggregate. Requirements should therefore be tied to decisions and risks, not abstract data perfection.

7.3 Interoperability and shared meaning

Interoperability has at least three dimensions. Technical interoperability allows systems to exchange messages. Syntactic interoperability ensures that message structures can be interpreted. Semantic interoperability ensures that exchanged terms have the same meaning. A diagnosis code, laboratory unit, medication identifier, or allergy concept must be governed consistently across systems.

Interfaces should preserve patient identity, encounter context, timestamps, authorship, status, and provenance. Interface monitoring is essential because a technically successful message may still be clinically unusable if it is mapped to the wrong patient, delayed, truncated, or displayed without its status.

8. Bottlenecks, Dependencies, and Process Redesign

8.1 Finding bottlenecks

A bottleneck is a constraint that limits throughput or creates accumulating delay. Symptoms include queues, repeated work, staff overloading, unfinished orders, delayed results, and frequent escalation. Analysis should distinguish a true capacity constraint from a coordination or information problem.

A simple flow relationship is often useful. If a process has average throughput λ patients per unit time and average time in the process W, then the average number of patients or cases in the process L is related by Little’s Law:

Here, L is measured in cases, λ in cases per unit time, and W in units of time. The relationship assumes a stable system over the period considered, with consistent definitions of entry and exit. For example, if an outpatient referral process completes 12 referrals per day and the average referral remains in the process for 3 days, the average work in progress is approximately 36 referrals. This does not identify the cause of delay; it helps quantify the relationship between backlog, throughput, and elapsed time.

Other useful measures include waiting time, processing time, queue length, first-pass yield, rework rate, cancellation rate, and percentage of orders requiring clarification. Measures should be stratified where appropriate by urgency, service, patient group, and time of day.

8.2 Dependencies and failure modes

A dependency exists when one activity cannot proceed, or cannot proceed safely, without another activity or information item. Examples include:

• medication administration depends on an authorized order and patient identity;
• surgery depends on consent, preoperative assessment, equipment, and scheduling;
• discharge depends on clinical readiness, medicines, transport, education, and follow-up;
• billing depends on coding, documentation, payer rules, and authorization.

Failure-mode analysis asks what can go wrong, why it may occur, how severe the consequence is, how likely it is, and how likely it is to be detected before harm. It should include technology failure, human factors, organizational conditions, and communication breakdowns.

8.3 Redesign principles

Effective redesign generally follows these principles:

1. Begin with the patient and clinical objective, not the software screen.
2. Remove unnecessary steps and duplicate data entry.
3. Capture information once and reuse it safely.
4. Place information and decision support at the point of work.
5. Standardize high-risk, repetitive processes while preserving professional discretion.
6. Design explicit exception paths rather than forcing workarounds.
7. Make responsibility, escalation, and pending work visible.
8. Test changes with real users in realistic conditions.
9. Measure unintended consequences, not only intended gains.
10. Implement incrementally with governance and feedback.

9. Illustrative Case: Redesigning the Emergency-to-Inpatient Journey

The following is a fictional illustrative case.

Background

Riverside General Hospital experienced long emergency-department stays for patients requiring admission. Clinicians reported that beds were unavailable even when the bed-management system showed vacancies. Nurses frequently repeated medication histories, and inpatient teams sometimes received incomplete handovers.

Problem

A multidisciplinary mapping exercise followed 30 representative journeys. The map showed four major problems: registration created duplicate patient records; admission requests lacked standardized clinical and isolation information; bed assignment was recorded in a separate system; and medication reconciliation began only after the patient reached the ward. The apparent bed shortage was partly a visibility and coordination problem.

Analysis

The team identified the following dependencies:

• bed assignment depended on a clinically complete admission request;
• transport depended on both bed confirmation and patient readiness;
• ward preparation depended on isolation and equipment requirements;
• medication reconciliation depended on access to community and prior hospital records;
• inpatient handover depended on a current problem list, pending results, and escalation plan.

Process data showed that the largest delays occurred after the admission decision but before physical transfer. A dashboard measured total emergency length of stay but did not identify the time spent waiting for clinical information, bed assignment, transport, or medication review.

Approach and solution

The hospital introduced a shared admission workflow with: a verified identity check; a structured admission request; mandatory but limited safety fields; real-time bed status; visible ownership of pending actions; and an electronic handover containing diagnosis, acuity, allergies, lines and devices, pending tests, medication issues, and escalation instructions. Pharmacy staff began medication reconciliation during the emergency stay for high-risk patients. The redesign was piloted on one medical unit, with daily review of exceptions.

Outcome

After the pilot, the hospital observed fewer duplicate records, fewer incomplete handovers, and better visibility of bed readiness. Total emergency stay improved modestly rather than dramatically, because transport and staffing constraints remained. This result was important: information redesign removed coordination waste but could not substitute for physical capacity or workforce planning.

Learning points

1. End-to-end mapping exposes delays hidden by departmental measures.
2. A shared information view is useful only when responsibilities and actions are explicit.
3. Mandatory fields should be limited to information that is necessary for safety or workflow progression.
4. Process redesign should distinguish information constraints from capacity constraints.
5. Pilot evaluation must include unintended effects, user workload, and patient safety.

10. Governance, Ethics, and Contemporary Developments

Health information is sensitive personal information. Process design must respect confidentiality, lawful access, purpose limitation, minimum necessary use, retention requirements, patient rights, and auditability. Role-based access should reflect professional responsibilities, while emergency access procedures should be available and reviewed. Consent requirements vary by jurisdiction and purpose; organizations must align system configuration with applicable law and policy.

Data quality and algorithmic fairness are also governance issues. If historical data underrepresent certain populations, an analytics or artificial-intelligence tool may produce less reliable recommendations for them. Automated prioritization should be clinically validated, monitored for unequal performance, and accompanied by human oversight. Users must understand whether an AI output is advisory, how it was generated at a practical level, and how to challenge or override it.

Contemporary HIS development increasingly involves cloud services, mobile documentation, patient portals, interoperability platforms, remote monitoring, process mining, and AI-assisted summarization. These technologies are valuable only when embedded in safe workflows. Process mining can discover actual sequences from event logs and compare them with intended pathways, but incomplete or inconsistent timestamps can create misleading maps. AI-generated summaries can reduce review time but must preserve source provenance and should not silently introduce unsupported conclusions.

Best-Practice Checklist

Before approving a workflow or HIS change, ask:

• Has the entire patient journey been mapped?
• Are all actors, handoffs, exceptions, and waiting states visible?
• Is each information element linked to a decision or action?
• Is the authoritative source and data owner defined?
• Are identity, status, time, provenance, and audit requirements addressed?
• Have safety, privacy, equity, and accessibility implications been assessed?
• Have frontline users and patients participated in design and testing?
• Are baseline and post-implementation measures defined?
• Are downtime, recovery, and escalation procedures clear?
• Has the organization planned for training, support, monitoring, and iterative improvement?

Chapter Summary

Health-system processes connect patient needs, professional work, information, resources, and organizational decisions. Mapping the patient journey provides an end-to-end perspective that complements departmental workflow analysis. Clinical, nursing, diagnostic, pharmacy, administrative, and managerial processes each have distinct information requirements, but they are linked through identity, encounters, orders, results, medication records, responsibilities, and handoffs.

Effective HIS design depends on understanding when information is created, who needs it, what action it supports, and what risks arise when it is inaccurate, late, incomplete, or inaccessible. Swim-lane maps, dependency analysis, bottleneck analysis, information-quality assessment, and requirements engineering provide practical tools for this work. Little’s Law can relate throughput, time, and work in progress when its steady-state assumptions are appropriate, but quantitative measures must be interpreted alongside clinical context.

Process redesign should reduce duplication, make pending work visible, support safe standardization, preserve professional judgment, and address exceptions. Governance must protect confidentiality, ensure accountability, and monitor the fairness and safety of analytics and AI. Ultimately, a hospital information system should be judged not by the number of screens or functions it contains, but by whether it helps people deliver safer, more coordinated, timely, equitable, and patient-centred care.

Key Terms

Process: A set of activities that transforms inputs into outputs.

Workflow: The sequence of activities, responsibilities, decisions, and information exchanges in work.

Patient journey: The patient’s experience across encounters, departments, transitions, and follow-up.

Care pathway: An agreed pattern of care for a condition or patient group.

Swim-lane map: A process map that assigns activities to actors or organizational units.

Handoff: Transfer of responsibility and relevant information between professionals or teams.

Information requirement: A defined need for information, including content, timing, user, format, and quality.

Data provenance: The origin, history, and transformations associated with data.

Master patient index: A system or process for identifying and linking patient records.

Clinical decision support: Functions that provide patient-specific information or recommendations to assist decisions.

Medication reconciliation: Comparison and resolution of medication information across transitions of care.

Diagnostic turnaround time: Elapsed time between defined diagnostic process points, such as order and verified result.

Bottleneck: A constraint that limits process capacity or causes accumulating delay.

Dependency: A relationship in which one activity relies on another activity, resource, or information item.

Interoperability: The ability of systems to exchange and meaningfully use information.

Semantic interoperability: Preservation of shared meaning across systems.

Information quality: The degree to which information is accurate, complete, timely, valid, consistent, and usable for its purpose.

Process mining: Analysis of event-log data to discover and evaluate actual workflows.

Role-based access control: Restriction of system access according to defined roles and permissions.

Work in progress: Cases that have entered a process but have not yet exited it.

Review Questions

Short Answer Questions

1. Distinguish a patient journey from a departmental workflow.
2. State four information-quality dimensions relevant to hospital systems.
3. What is the purpose of a swim-lane process map?
4. Why is medication reconciliation important at transitions of care?
5. Give two examples of dependencies in a hospital process.

Long Answer Questions

1. Explain how clinical, nursing, diagnostic, and pharmacy workflows interact during an inpatient admission.
2. Discuss the functional and non-functional requirements of a hospital information system.
3. Explain how poor patient identity management can affect clinical, administrative, and managerial processes.
4. Evaluate the role of handoff design in patient safety and continuity of care.
5. Discuss the benefits and limitations of process redesign based on digital technologies.

Analytical / Application Questions

1. Select an outpatient journey and construct a swim-lane map showing actors, information exchanges, waiting points, and exception paths. Recommend two redesign interventions.
2. A service completes 20 referrals per day and the average referral remains open for 2.5 days. Using Little’s Law, estimate the average work in progress and explain the assumptions behind the estimate.
3. A hospital introduces an AI tool to prioritize diagnostic work. Design an evaluation plan addressing accuracy, timeliness, fairness, clinician oversight, privacy, and unintended consequences.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a patient journey from a departmental workflow?

A. A patient journey covers the experience across connected services, while a workflow describes activities within a process.

B. A patient journey contains only clinical information, while a workflow contains only financial information.

C. A patient journey is always linear, while a workflow is always circular.

D. A patient journey is used only for billing, while a workflow is used only for diagnosis.

Correct Answer: A. A patient journey covers the experience across connected services, while a workflow describes activities within a process.

Question 2

Which information is most directly required to support safe medication administration?

A. The hospital’s annual pharmacy budget

B. The authorized medicine order linked to the correct patient, dose, route, and time

C. The patient’s appointment wait time from the previous year

D. The laboratory department’s staffing roster

Correct Answer: B. The authorized medicine order linked to the correct patient, dose, route, and time

Question 3

In a swim-lane map, lanes primarily represent:

A. Different clinical diagnoses

B. Different data-storage formats

C. Actors or organizational units responsible for activities

D. Different levels of disease severity

Correct Answer: C. Actors or organizational units responsible for activities

Question 4

Which is an example of semantic interoperability?

A. Two systems use the same network cable

B. Two systems exchange data in different file sizes

C. Two systems display information on different screen types

D. Two systems interpret a laboratory code and unit with the same meaning

Correct Answer: D. Two systems interpret a laboratory code and unit with the same meaning

Question 5

A process completes 12 referrals per day and the average referral remains open for 3 days. Under steady-state assumptions, what is the approximate average work in progress?

A. 4 referrals

B. 9 referrals

C. 36 referrals

D. 48 referrals

Correct Answer: C. 36 referrals

Question 6

Which feature most directly supports safe diagnostic result management?

A. The ability to overwrite all previous results without a history

B. A distinction between preliminary, final, corrected, and amended results

C. Removal of specimen collection times

D. Automatic release of every result without validation

Correct Answer: B. A distinction between preliminary, final, corrected, and amended results

Question 7

What is a major risk of excessive clinical decision-support alerts?

A. Improved semantic interoperability

B. Alert fatigue and reduced attention to important alerts

C. Elimination of all documentation requirements

D. Guaranteed reduction in medicine costs

Correct Answer: B. Alert fatigue and reduced attention to important alerts

Question 8

Which measure best reflects information quality?

A. Whether information is accurate, timely, complete, and fit for the decision being made

B. The number of screens available in the system

C. The age of the hardware used by the hospital

D. The number of employees with email accounts

Correct Answer: A. Whether information is accurate, timely, complete, and fit for the decision being made

Question 9

In the illustrative emergency-to-inpatient case, which problem contributed to delay after the admission decision?

A. The absence of any clinical staff

B. The lack of a shared view of bed readiness and pending actions

C. The complete absence of diagnostic services

D. The use of paper medicine packaging

Correct Answer: B. The lack of a shared view of bed readiness and pending actions

Question 10

Which is an appropriate governance measure for an AI-supported prioritization tool?

A. Allowing the tool to make unreviewable decisions

B. Assessing performance across patient groups and retaining human oversight

C. Deleting the data used to produce each recommendation

D. Disabling audit logs to improve speed

Correct Answer: B. Assessing performance across patient groups and retaining human oversight