Integrated Care Operations Evidence Report

1. Purpose of this Report

This report consolidates and rewrites the evidence supplied in three source documents into a single, deduplicated OmniSenti report. The task was not to conduct a fresh evidence search, but to convert the supplied evidence into a usable market research report that answers ten practical questions about hospitals, aged care facilities, healthcare ERP, interoperability and AI.

The report deliberately avoids naming or directly promoting any specific platform. Its purpose is to support strategic market communication by setting out the independent evidence for a simple proposition: fragmented healthcare operations carry measurable cost, risk and staff-time penalties, while integrated operational systems and selected AI tools can reduce those penalties when implemented with sound governance.

2. Evidence Handling and Deduplication Method

The three supplied evidence documents contained overlapping material. Repeated evidence has been consolidated so that the same study or statistic is generally cited once as a source item, then reused through a common source code where relevant. For example, the Australian medication-risk evidence from Griffith University and the Australian Commission medication-error evidence appeared in more than one source document, but they are treated once in the reference register.

Evidence was prioritised in the following order:

Peer-reviewed academic studies.
Government reports and public agencies.
Professional bodies and recognised public-interest bodies.
Independent research organisations and industry analyses where the methodology was visible.
Supplier or vendor-published material was excluded from the primary evidence base unless clearly labelled as lower-confidence implementation evidence. The main analysis does not rely on software-seller claims.

Several questions do not have a single official benchmark for a "typical" 100-bed aged care facility or hospital. Where that is the case, this report provides a documented range or a model derived from published data. The calculations are planning estimates, not guaranteed savings.

3. Executive Summary

The independent evidence supports ten broad findings.

Question area	Evidence-backed answer
Medication-error cost	A hospital planning proxy is approximately USD 400,000 per 100 beds per year for preventable adverse drug events. Australian national data implies around AUD 5,600 per medication-related admission, while aged-care medication information risks support a proxy of about AUD 153,000 per 100 residential aged care users per year.
Staff hours spent on administration	Healthcare staff commonly spend between 20% and 35% of time on documentation or administration-heavy work. Physicians spending 49.2% of office-day time on EHR and desk work, hospital nurses spending 35.3% on documentation, and Australian nursing-home enrolled nurses spending 29.0% on documentation plus 37.7% on communication.
Duplicate data entry and repeated entry	The most defensible practical benchmark is 25 to 30 minutes per clinician per day on referral/follow-up rework. For 20 clinicians this equates to 42 to 50 hours per week on a five-day basis, or 58 to 70 hours per week across seven-day operations.
Disconnected systems	Evidence supports a planning range of 10 to 20 partly connected systems or device classes for a mid-size hospital or aged-care provider. One hospital case integrated 12 legacy departmental systems, while OECD evidence reports an average of around 12 hospital eMR vendors across surveyed countries.
Limited expense visibility	US healthcare waste is estimated at USD 760bn to USD 935bn annually. Hospital supply expenses average roughly 15% of hospital expenses, with 17.4% reducible supply expense and USD 12.1m average annual opportunity per hospital. Australian digital-health productivity work estimates potential savings above AUD 5bn per year.
Fragmented patient records	Physicians spent 16 minutes 14 seconds per outpatient encounter in EHRs, with 33% on chart review. A primary-care study found 36.2 minutes of EHR time per visit, including inbox and after-hours work.
Healthcare ERP evidence	ERP-specific healthcare studies exist but are often case-study or perception-based. The strongest quantified evidence comes from ERP-relevant functions: eMAR, barcode medication administration, CPOE, interoperability, documentation tools, supply-chain visibility and finance integration.
100-bed interconnectivity saving	A practical evidence-based planning range is 60 to 200 staff hours per week for a 100-bed facility. Higher scenarios, up to around 275 hours per week, are plausible only where documentation burden and fragmentation are severe.
Medication-administration risk reduction	Documented results include 4.2 fewer medication administration errors per 100 administrations, a 56% decline in potentially serious medication administration errors, and reduction of potentially harmful medication administration errors from 3.0% to 0.3% in one barcode-assisted study.
AI time and money savings	Studies report 13 fewer EHR minutes and 16 fewer documentation minutes per clinician per day, or 15% documentation-time reduction in another study. For 20 clinicians, that supports a planning estimate of roughly 30 to 37 hours per week saved from AI scribe use alone.

4. Question-by-Question Findings

Question 1. Using a typical aged care facility or hospital, what is the average annual cost of medication errors?

Answer

There is no single accepted average cost for medication errors in a typical hospital or aged care facility. However, the supplied evidence supports three practical cost lenses:

Hospital adverse drug event cost: A classic US hospital study estimated annual costs of USD 5.6 million for all adverse drug events and USD 2.8 million for preventable adverse drug events in a 700-bed teaching hospital. Scaled linearly, this is approximately USD 800,000 per 100 beds for all adverse drug events and USD 400,000 per 100 beds for preventable adverse drug events before inflation [S1].
Australian medication-related admissions: The Australian Commission on Safety and Quality in Health Care reported approximately 250,000 medication-related hospital admissions per year in Australia, costing about AUD 1.4 billion, implying roughly AUD 5,600 per medication-related admission [S3].
Australian aged-care medication information risk: Griffith University research estimated medication information risks in aged care at approximately AUD 312 million annually. Using the national residential aged-care user denominator of about 204,000 residents, this equates to roughly AUD 153,000 per 100 residential aged care users per year as a system-level proxy [S4], [S6].

Evidence table

Evidence source	Key documented finding	Facility-level interpretation
Bates et al., JAMA	USD 5.6m annual all-ADE cost and USD 2.8m preventable-ADE cost in a 700-bed teaching hospital [S1].	Approx. USD 800k per 100 beds for all ADEs and USD 400k per 100 beds for preventable ADEs, before inflation.
Elliott et al., BMJ Quality & Safety	England: definitely avoidable adverse drug events costing the NHS about GBP 98.5m and consuming 181,626 bed-days [S2].	Facility cost is driven by avoidable bed-days, admissions, escalation and harm events.
Australian Commission on Safety and Quality in Health Care	250,000 medication-related admissions costing AUD 1.4bn annually [S3].	About AUD 5,600 per medication-related admission.
Griffith University	Medication information risks in aged care estimated at AUD 312m annually [S4].	Approx. AUD 153k per 100 residential aged care users as a national-allocation proxy.
Silva et al.	Residential aged-care medication incident rate of 1.81 per 1,000 resident-days [S5].	A 100-resident facility at full occupancy would expect about 66 medication incidents per year.

Practical conclusion

A conservative strategic range for a 100-bed or 100-care-recipient facility is AUD/USD 150,000 to 400,000 per year in medication-error or medication-risk exposure, depending on whether the facility resembles aged care, subacute care or acute hospital care. This should not be presented as a guaranteed saving. It is a risk-cost planning range requiring local baseline validation.

Evidence confidence: High for system-level and hospital costs. Medium for translating national aged-care costs into a typical facility estimate.

Question 2. Using a typical aged care facility or hospital, what is the average percentage of staff hours spent on administrative tasks?

Answer

The evidence shows that documentation and administration-heavy work consumes a substantial share of clinical labour. A defensible planning range is 20% to 35% of staff time in many hospital or aged-care settings, with some roles and environments higher.

US physicians spent 49.2% of office-day time on EHR and desk work and 27.0% on direct clinical face time [S8].
A 36-hospital nursing time-motion study found documentation was the largest observed category, at 147.5 minutes, or 35.3% of nurses' time [S9].
AHRQ evidence reports ICU documentation falling from 35.1% to 24.2% after implementation of an ICU information system [S10].
A systematic review found nurses commonly spent 13% to 31% of time on documentation [S11].
Australian nursing-home research found enrolled nurses spent 29.0% of time on documentation and 37.7% on communication [S12].
Swiss nursing-home evidence found around one-third of care workers and nearly half of registered nurses spent two or more hours per normal day on administrative activities [S13].

Facility interpretation

A typical 100-bed aged-care facility can be modelled using 100 residents and 215 care minutes per resident per day, giving about 358 care hours per day [S7]. If documentation and administration-related work consumes 20% to 35% of this time, then approximately 72 to 125 staff hours per day are being spent on documentation, administration, communication, searching, re-entry and related work.

Not all of that time is waste. Some documentation is clinically or legally necessary. The commercial improvement opportunity lies in the avoidable portion: duplicate entry, searching for missing information, repeated form completion, chasing approvals, reconciling inconsistent records and manually stitching systems together.

Evidence confidence: High for the existence and scale of administrative/documentation burden. Medium for transferring exact percentages into a specific facility without local measurement.

Question 3. Using a typical aged care facility or hospital, what is the average monthly or weekly time wasted on duplicate data entry?

Answer

Direct measurements of duplicate data entry are less common than broader measurements of documentation and EHR burden. The clearest directly usable benchmark in the supplied evidence is the Royal College of General Practitioners finding that most GP participants spent 25 to 30 minutes per day on referral and follow-up tasks involving manual data entry of non-integrated forms, re-issuing prescriptions and re-sending referrals that were lost, bounced back or rejected [S14].

Unit	Estimated duplicate-entry / rework burden
Per clinician	2.1 to 2.5 hours per week, or 9 to 11 hours per month.
20 clinicians, five-day basis	42 to 50 hours per week, or about 180 to 220 hours per month.
20 clinicians, seven-day operating basis	58 to 70 hours per week.
50 clinicians, five-day basis	105 to 125 hours per week, or about 450 to 550 hours per month.

Supporting evidence

Poissant et al.'s systematic review found that EHRs using bedside terminals and central-station desktops saved nurses 24.5% and 23.5% of documentation time respectively [S15]. AHRQ evidence also found ICU documentation time fell from 35.1% to 24.2% after an ICU information system [S10].

A duplicate medical record study reported a 4% monthly duplicate record creation rate, reduced to 2.3% after usability redesign and sustained at around 2.5% over two years [S44].

Practical conclusion

For market sizing and operational planning, use 40 to 70 staff hours per week as a conservative duplicate-entry and system-rework range for a 100-bed facility with around 20 clinicians or senior care users of the system.

Evidence confidence: Medium. The activity is real and documentable, but it is often measured inside broader categories.

Question 4. Using a typical aged care facility or hospital, what are the typical number of disconnected or not entirely integrated software or hardware systems?

Answer

No single official average was identified. However, the independent evidence supports a realistic planning range of 10 to 20 partially connected systems or device classes for a mid-size healthcare organisation.

Evidence source	Finding	Interpretation
Cruz-Correia et al.	A hospital virtual EPR integration project connected data from 12 legacy departmental information systems plus administrative data [S18].	A real hospital example demonstrates double-digit legacy-system fragmentation.
OECD	Surveyed countries reported average numbers of available eMR vendors of about 16 for physicians, 22 for specialists and 12 for hospitals [S16].	Healthcare digital ecosystems are vendor-fragmented.
ONC / NCBI Data Brief	In 2023, 70% of US hospitals engaged in all four domains of interoperability: send, receive, find and integrate [S17].	Interoperability is improving but not universal.
Digital hospital downtime evidence	A 350-bed digital hospital experienced downtime events disrupting integrated EMR-related care delivery [S19].	Integration creates dependency on resilience and governance.

Typical system and device categories

Aged care facilities and hospitals commonly operate separate or partly integrated systems for:

Patient or resident administration.
Clinical record or care management.
Medication chart, eMAR or medication management.
Pharmacy or dose-administration interface.
Rostering and workforce scheduling.
Payroll or HR.
Billing, claims or patient accounts.
Finance and accounting.
Inventory, procurement or supply chain.
Incident and compliance management.
Document management or scanning.
Laboratory, radiology or referral systems.
SMS, email or communications.
Reporting or business intelligence.
Nurse call, vitals devices, infusion devices, pumps, telehealth or other medical hardware.

Practical conclusion

A practical estimate for a typical 100-bed facility is 10 to 20 disconnected or semi-integrated systems or device classes. More complex hospital groups may exceed that range. The important issue is not only the count of systems, but the number of workflow handoffs requiring staff to manually re-enter, reconcile or chase information.

Evidence confidence: Medium. Fragmentation is well documented; an exact universal system count is not.

Question 5. What metrics demonstrate that limited visibility into expenses leads to costly inefficiencies, and what are the figures?

Answer

Limited expense visibility is best measured through waste, administrative complexity, supply-chain spend, inventory leakage, delayed finance insight and missed reconciliation. The strongest quantified evidence comes from US healthcare waste studies, supply-chain studies and Australian digital-health productivity analysis.

Evidence source	Metric	Figure
Shrank et al., JAMA	Estimated annual waste in the US healthcare system [S21].	USD 760bn to USD 935bn.
Shrank et al., JAMA	Administrative complexity in the US healthcare waste taxonomy [S21].	USD 265.6bn.
Abdulsalam et al.	Hospital supply expenses as a share of total hospital expenses [S23].	Average around 15%; can reach 30% to 40% in high case-mix hospitals.
Navigant / Guidehouse research reported by Healthcare Dive	Large-hospital supply-chain opportunity across 2,127 hospitals [S24].	USD 12.1m average annual opportunity per hospital; 17.4% reducible supply expense.
Productivity Commission, Australia	Potential national saving from better digital healthcare integration [S22].	More than AUD 5bn per year.

100-bed facility model

Annual cost opportunity = facility operating expense x supply spend share x avoidable inefficiency percentage

Example facility operating expense	Supply spend share	Avoidable inefficiency	Annual opportunity
AUD 20m	15%	5%	AUD 150,000
AUD 20m	15%	10%	AUD 300,000
AUD 20m	15%	17.4%	AUD 522,000

Practical conclusion

For an AUD 20m annual operating-cost base, AUD 150,000 to AUD 522,000 per year is a plausible expense-visibility opportunity from better supply, inventory and finance control alone.

Evidence confidence: High for system-level waste and supply-chain figures. Medium for facility-level extrapolation.

Question 6. What metrics demonstrate that fragmented patient records and related files cause staff to lose time, and how much time is lost?

Answer

Fragmented patient records create time loss through searching, chart review, repeated referral activity, duplicate form completion, lost referrals, bounced-back referrals, inbox management, reconciliation and after-hours EHR work.

Evidence source	Finding	Time-loss interpretation
RCGP	Most GP participants spent 25 to 30 minutes per day on referral and follow-up tasks involving manual data entry, re-sending or rejected referrals [S14].	Direct evidence of fragmentation-related rework.
Overhage et al.	Physicians spent 16 minutes 14 seconds per outpatient encounter using the EHR; chart review was 33% of that time [S25].	About 5 minutes 22 seconds per encounter is chart review, highly affected by record design and completeness.
Rotenstein et al.	Primary care physicians spent a median of 36.2 minutes of EHR time per visit, including inbox and after-hours work [S26].	High record burden creates a large target for improved information architecture and automation.
Iscoe et al.	Emergency physicians spent a median of 6.82 minutes of EHR time per encounter [S27].	Even high-throughput emergency care has measurable EHR time per encounter.
Hersh et al.	HIE evidence suggests reductions in duplicative lab/radiology testing, ED costs and admissions [S28].	Interoperability can reduce repeated tests, unnecessary retrieval and avoidable downstream work.

Facility-level estimate

Using the RCGP benchmark alone, a 20-clinician facility loses around 42 to 50 hours per week if each clinician loses 25 to 30 minutes per working day across five days. If the same burden occurs across seven-day operations, the estimate becomes 58 to 70 hours per week.

Practical conclusion

A reasonable conservative planning figure is 40 to 70 staff hours per week in fragmented-record rework for a 100-bed facility with around 20 regular clinical system users.

Evidence confidence: High that fragmentation causes time loss. Medium for precise per-facility hours.

Question 7. What documented and verifiable reports demonstrate how implementation of healthcare-specific ERP systems improved efficiency across operations?

Answer

The evidence base for "healthcare ERP" is more mixed than the evidence base for specific operational capabilities. Independent ERP literature in healthcare often consists of case studies, implementation analyses or perception studies rather than controlled ROI trials. Stronger quantified evidence comes from the operational modules that a healthcare ERP should contain: medication management, ordering, eMAR, barcode medication administration, EHR documentation, interoperability, supply-chain visibility and finance integration.

ERP-specific evidence

Evidence source	Type	Finding and relevance
Agarwal et al.	Hospital ERP case study [S29].	Describes implementation phases, process flow and stakeholder issues in hospital ERP implementation. Useful for delivery planning and organisational readiness.
Fiaz et al.	Mixed-methods ERP service-quality study [S30].	Found ERP-related information quality, system quality and organisational factors influenced perceived healthcare service quality.
Engelmann et al.	ERP expansion toward real-time operation planning [S31].	Shows that expanding ERP scope can support real-time operational control.
Journal of Enterprise Information Management hospital ERP case	Hospital SAP R/3 integration case [S32].	Describes ERP integration issues in a hospital environment.

Stronger quantified ERP-relevant evidence

ERP-relevant capability	Documented effect
Electronic medication administration records	Australian evidence briefing reports reduced medication administration errors such as dose omissions and timing errors, improved medication documentation and improved quality indicators [S33].
Barcode medication administration	Poon et al. found bar-code eMAR substantially reduced order transcription errors, medication administration errors and potential ADEs [S35].
Automated unit-dose dispensing with barcode-assisted medication administration	Overall medication administration errors fell from 19.5% to 15.8%, and potentially harmful errors fell from 3.0% to 0.3% [S36].
Documentation and EHR efficiency	Poissant et al. found reductions in documentation time, including nurse time savings of around 23% to 25% in relevant settings [S15].
Interoperability and HIE	Evidence suggests reductions in duplicative tests, ED costs and admissions, though the quality and consistency of evidence vary [S28].
Supply-chain visibility	Hospital supply expense studies show large savings opportunities when supply spend is visible and controlled [S23], [S24].

Practical conclusion

It is more accurate to state that healthcare-specific ERP improves efficiency when it integrates proven operational modules into one workflow and data spine, rather than claiming that an ERP label alone creates savings. The value case should be framed as integration of patient identity, scheduling, medication workflow, workforce eligibility, billing, finance, inventory, reporting and audit.

Evidence confidence: Medium for ERP-specific quantified outcomes. High for module-specific operational outcomes that a healthcare ERP should integrate.

Question 8. In a typical 100-bed aged care or hospital facility, how many staff hours can be saved through improving interconnectivity between disparate systems?

Answer

A realistic evidence-based planning range is 60 to 200 staff hours per week for a 100-bed or 100-care-recipient facility, depending on staffing model, baseline documentation burden and system fragmentation. Higher estimates, up to around 275 hours per week, are plausible where documentation burden is high and many workflows are duplicated.

Evidence-based scenarios

Scenario	Evidence base	Calculation	Estimated saving
Conservative duplicate-entry scenario	RCGP: 25 to 30 minutes per clinician per day on referral/follow-up and non-integrated forms [S14].	20 clinicians x 25 to 30 minutes/day x 7 days.	58 to 70 hours/week.
Moderate documentation-efficiency scenario	Poissant et al.: EHR documentation saved about 24% of documentation time [S15].	100 residents x 215 minutes/day = 358 care hours/day. If 20% to 35% is documentation and 24% is saved.	120 to 211 hours/week.
High information-system scenario	AHRQ: ICU documentation fell from 35.1% to 24.2%, saving 10.9 percentage points [S10].	10.9% x 358 care hours/day x 7 days.	About 273 hours/week.

Practical conclusion

For external communication, the safest figure is 60 to 200 staff hours per week of opportunity in a 100-bed care environment. The upper end should only be used after local baselining confirms severe duplication, high documentation burden or multiple manual handoffs.

Evidence confidence: Medium. This is a modelled estimate based on published time-motion and workflow evidence.

Question 9. Precisely how has implementation of healthcare-specific ERP systems reduced medication-administration risk indicators?

Answer

The strongest evidence does not show that a generic ERP automatically reduces medication-administration risk. It shows that ERP-connected medication controls reduce risk when the operating platform includes eMAR, electronic prescribing or ordering, barcode medication administration, closed-loop pharmacy workflow, patient identity controls, medication reconciliation, role controls, audit and alerts.

Risk indicators that can be reduced

A healthcare ERP or integrated care-operations system can reduce and track the following medication-administration risk indicators:

Wrong patient.
Wrong medication.
Wrong dose.
Wrong route.
Wrong time.
Omitted dose.
Late dose.
Unsigned administration.
Medication administered without current order.
Allergy or interaction override without review.
Controlled-drug discrepancy.
Non-scanned patient wristband.
Non-scanned medication.
Transition-of-care medication discrepancy.
Missing medication reconciliation.
Missing administration reason for omitted dose.
Duplicate medication task.
Missed follow-up for high-risk medication.

Documented reductions

Evidence source	Documented effect
Australian Commission eMAR evidence briefing	Good-quality evidence that eMAR systems are associated with reduced medication administration errors such as dose omissions and timing errors, improved quality indicators and enhanced medication documentation [S33].
Australian Commission eMAR evidence briefing	The largest controlled before-after Australian study reported a reduction of 4.2 medication administration errors per 100 administrations and a 56% decline in potentially serious medication administration errors [S33].
Jessurun et al.	Overall medication administration errors fell from 19.5% to 15.8% after automated unit-dose dispensing with barcode-assisted medication administration; potentially harmful errors fell from 3.0% to 0.3% [S36].
Poon et al.	Barcode eMAR substantially reduced transcription errors, medication administration errors and potential adverse drug events [S35].
Australian Commission closed-loop medication briefing	Closed-loop medication systems can reduce dispensing and administration errors and reduce medication turnaround time, with implementation and workflow design being critical [S34].

Mechanism of risk reduction

Healthcare ERP reduces medication-administration risk only when it changes workflow. The core mechanisms are:

Single patient identity: reduces patient-record mismatch and duplicate patient risks.
Medication task linkage: connects medication orders, administration tasks, patient record and staff workflow.
Role and credential controls: restricts medication administration tasks to authorised staff.
Barcode or scan workflow: confirms patient, medication, dose and timing at administration.
eMAR audit: records due, overdue, omitted, refused and administered states.
Medication reconciliation: reduces transition-of-care discrepancies.
Pharmacy integration: aligns medication changes, supply and administration records.
Exception work queues: surfaces missing orders, overdue doses, rejected chart changes and high-risk medicines.
Risk reporting: tracks risk indicators by site, ward, unit, provider, medication class and workflow stage.

Evidence confidence: High for eMAR, barcode medication administration and closed-loop medication systems. Medium for attributing the effect to "ERP" unless the ERP includes those controls.

Question 10. In a typical 100-bed aged care facility or hospital, what documented savings in time or money have resulted from utilising AI, and what specific AI solution was implemented?

Answer

The strongest independent evidence is currently for ambient clinical documentation tools and acute-care prediction tools. Aged-care-specific AI ROI evidence is emerging but is thinner and often supplier-published. This report therefore treats AI savings carefully.

Independent evidence table

AI solution	Evidence source	Documented outcome	100-bed interpretation
Ambient AI scribe / documentation assistant	Rotenstein et al., JAMA Network Open / PubMed	Associated with 13 fewer EHR minutes and 16 fewer documentation minutes per clinician per day, plus modest increases in weekly visits [S39].	For 20 clinicians, approx. 4.3 to 5.3 hours/day, or 30 to 37 hours/week.
Ambient AI scribe	Tan et al., JMIR Medical Informatics	Documentation time reduced by 15%, from mean 5.3 to 4.5 minutes per consultation [S38].	At 150 consultations/day, saving is about 120 minutes/day, or 14 hours/week.
Ambient AI scribe / documentation burden	Olson et al.	Quality-improvement evidence indicates ambient AI scribes reduce documentation burden [S37].	Useful for clinician documentation relief in consult-heavy services.
Sepsis prediction algorithm	Burdick et al.	Reported 39.5% mortality reduction, 32.3% length-of-stay reduction and 22.7% 30-day readmission reduction after implementation [S40].	Applies mainly to acute hospitals with sepsis-risk populations, not directly to low-acuity aged care.
Severe sepsis / septic shock prediction algorithm	Calvert et al.	For a 50-bed ICU, estimated 75 additional lives saved per year and USD 560k cost reduction from earlier prediction [S41].	Strong hospital/ICU evidence, but not a general 100-bed aged-care estimate.
AI scribe implementation governance	Australian Digital Health Agency	Provides workflow and implementation guidance for AI scribes in healthcare [S42].	Supports safe implementation planning, not a direct ROI figure.

Practical 100-bed AI savings scenarios

Scenario	AI use case	Assumption	Estimated staff-time saving
Conservative	Ambient AI scribe in consult-heavy service.	20 clinicians save 13 to 16 minutes per day.	30 to 37 hours/week.
Moderate	AI scribe plus referral/document generation.	20 clinicians save 20 to 30 minutes per day.	47 to 70 hours/week.
Acute-care high impact	AI scribe plus sepsis/deterioration AI.	Documentation savings plus outcome and length-of-stay effects.	Staff-time savings vary; cost benefits can be much larger but require hospital-specific baseline data.

Practical conclusion

For a typical 100-bed facility, the safest independent AI claim is that ambient documentation AI can save about 30 to 37 clinician hours per week where 20 clinicians use it consistently. Larger savings are possible when AI is combined with referral automation, care-plan generation, patient-flow prediction or clinical deterioration tools, but those require more careful local validation.

Evidence confidence: High for modest documentation-time savings from ambient AI scribes. Medium to high for sepsis algorithms in acute hospital settings. Low to medium for broad aged-care AI ROI until more independent aged-care implementation studies are available.

5. Consolidated Evidence Map

The table below lists the major evidence themes without repeating the same source multiple times.

Evidence theme	Main deduplicated source items
Medication-error cost	Bates [S1], Elliott [S2], Australian Commission [S3], Griffith [S4], Silva [S5].
Administrative burden	Sinsky [S8], Hendrich/Permanente [S9], AHRQ [S10], Lim [S11], Munyisia [S12], Ausserhofer [S13].
Duplicate entry and repeated work	RCGP [S14], Poissant [S15], duplicate medical-record usability study [S44].
Disconnected systems	Cruz-Correia [S18], OECD [S16], ONC / NCBI [S17], Chen [S19].
Expense visibility	Shrank [S21], Productivity Commission [S22], Abdulsalam [S23], Healthcare Dive / Navigant [S24].
Fragmented records	RCGP [S14], Overhage [S25], Rotenstein 2023 [S26], Iscoe [S27], Hersh [S28].
Healthcare ERP and ERP-relevant improvements	Agarwal [S29], Fiaz [S30], Engelmann [S31], hospital ERP case [S32], eMAR [S33], closed-loop medication [S34], Poon [S35], Jessurun [S36].
AI evidence	Olson [S37], Tan [S38], Rotenstein 2026 [S39], Burdick [S40], Calvert [S41], ADHA AI scribe guidance [S42].

6. Strategic Interpretation for Healthcare ERP and AI Programs

The evidence points to a practical market narrative for integrated healthcare operations.

6.1 Fragmentation is a measurable operating cost

The strongest evidence does not merely say that disconnected systems are inconvenient. It shows that staff lose time in repeated documentation, manual data entry, chart review, referral rework, system switching, record reconciliation, duplicate testing and manual follow-up. The cost is paid through staff hours, delays, risk exposure and poor finance visibility.

6.2 Medication risk is a workflow design problem

Medication safety is not only a matter of individual vigilance. Risk increases when patient identity, medication orders, pharmacy updates, administration tasks, staff credentials and transition information are disconnected. The evidence for eMAR, barcode administration and closed-loop medication workflow shows that system design can reduce medication-administration risk indicators.

6.3 ERP value comes from joining proven modules

Independent ERP-specific evidence exists, but the strongest numerical evidence sits in the modules that an ERP should integrate. A healthcare ERP value proposition is therefore strongest when it is framed as the integration layer for proven improvement levers: patient master data, scheduling, workforce eligibility, medication workflow, billing, finance, inventory, reporting and audit.

6.4 AI is credible when tied to specific workflows

AI claims are strongest when tied to concrete use cases: ambient documentation, structured note generation, referral drafting, care-plan generation, triage support and deterioration prediction. Broad claims that AI will transform an entire facility are less defensible unless tied to baseline measures and implementation scope.

7. Recommended Measurement Program

For any healthcare provider considering integrated ERP or AI implementation, OmniSenti recommends a baseline-and-after measurement program.

7.1 Baseline metrics before implementation

Capture for four to eight weeks:

Medication administration risk indicators per 1,000 resident-days or bed-days.
Dose omissions, late doses, unsigned administrations and refused administrations.
Number of manual medication chart changes.
Duplicate data entry events per staff member per shift.
Staff minutes spent on documentation, searching and re-entry.
Referral and external-message rework events.
Billing rework, claim rejection and adjustment rates.
Time to produce management, finance and compliance reports.
Number of systems used in each workflow.
Roster change frequency and time spent filling shifts.
Patient/resident record completeness.
Time spent reconciling inventory, expenses, claims or payments.

7.2 Post-implementation measurement points

Measure again at:

30 days;
90 days;
180 days;
12 months.

7.3 Suggested headline KPIs

Staff hours saved per week.
Medication risk indicators reduced per 1,000 resident-days.
Duplicate-entry events reduced.
Claim rejection rate reduced.
Time to close billing period.
Time to produce compliance report.
Roster fill time reduced.
Overdue medication-task rate reduced.
Patient/resident record completeness improved.
User adoption and task completion rate.

8. Evidence Limitations

Facility averages are rarely universal. Many studies are national, system-level, clinician-level or unit-level, not "one average 100-bed facility."
Aged-care cost data is thinner than hospital data. Medication risk and administrative burden are well documented, but dollar cost by aged-care facility is less robust.
ERP-specific ROI evidence is often indirect. Many healthcare ERP studies are case studies, perception studies or implementation analyses. The stronger numerical evidence sits in ERP-relevant modules.
AI evidence is maturing quickly. Ambient AI documentation evidence is comparatively strong, while broad AI automation ROI in aged care remains less independently established.
Local baselining is essential. Published evidence should justify the problem and shape the model. The provider's own baseline should prove the saving.

9. References and Source Links

[S1] Bates DW, Spell N, Cullen DJ, et al. "The costs of adverse drug events in hospitalized patients." JAMA. 1997. https://pubmed.ncbi.nlm.nih.gov/9002493/
[S2] Elliott RA, Camacho E, Jankovic D, et al. "Economic analysis of the prevalence and clinical and economic burden of medication error in England." BMJ Quality & Safety. 2021. https://qualitysafety.bmj.com/content/30/2/96
[S3] Australian Commission on Safety and Quality in Health Care. "National effort to cut medicine errors at transitions of care." 2025. https://www.safetyandquality.gov.au/newsroom/latest-news/national-effort-cut-medicine-errors-transitions-care
[S4] Griffith University. "Griffith research reveals medication information risks in aged care." 2026. https://news.griffith.edu.au/2026/03/05/griffith-research-reveals-medication-information-risks-in-aged-care/
[S5] Silva SSM, et al. "Characteristics and risk factors of medication incidents in residential aged care." 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12860823/
[S6] Australian Institute of Health and Welfare / GEN Aged Care Data. "People using aged care." https://www.gen-agedcaredata.gov.au/topics/people-using-aged-care
[S7] My Aged Care. "Nursing and personal care in aged care homes." https://www.myagedcare.gov.au/nursing-and-personal-care-aged-care-homes
[S8] Sinsky C, Colligan L, Li L, et al. "Allocation of physician time in ambulatory practice." Annals of Internal Medicine. 2016. https://pubmed.ncbi.nlm.nih.gov/27595430/
[S9] Hendrich A, Chow MP, Skierczynski BA, Lu Z. "A 36-hospital time and motion study: how do medical-surgical nurses spend their time?" The Permanente Journal. 2008. https://www.thepermanentejournal.org/doi/10.7812/tpp/08-021
[S10] Agency for Healthcare Research and Quality. "Impact of Health IT on Nurses' Time Spent on Direct Patient Care." https://digital.ahrq.gov/sites/default/files/docs/page/impact-of-hit-on-nurses-quick-reference-guide.pdf
[S11] Lim AG, et al. "Time-motion study of clinical nursing documentation." 2019. https://journals.sagepub.com/doi/10.1177/2010105819834569
[S12] Munyisia EN, Yu P, Hailey D. "How nursing staff spend their time on activities in a nursing home: an observational study." Journal of Advanced Nursing. 2011. https://pubmed.ncbi.nlm.nih.gov/21466577/
[S13] Ausserhofer D, et al. "Administrative burden experienced by nursing home staff." 2023. https://pmc.ncbi.nlm.nih.gov/articles/PMC10237523/
[S14] Royal College of General Practitioners. "Tackling the GP workload crisis: From evidence to action on hidden and avoidable workload in general practice." 2026. https://www.rcgp.org.uk/getmedia/38760b50-f2a2-43f0-9d1f-8986edb811e6/RCGP-Tackling-the-GP-workload-crisis-April-2026.pdf
[S15] Poissant L, Pereira J, Tamblyn R, Kawasumi Y. "The impact of electronic health records on time efficiency of physicians and nurses: a systematic review." Journal of the American Medical Informatics Association. 2005. https://pmc.ncbi.nlm.nih.gov/articles/PMC1205599/
[S16] OECD. "Progress on implementing and using electronic health record systems." 2023. https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/09/progress-on-implementing-and-using-electronic-health-record-systems_f6c2a59a/4f4ce846-en.pdf
[S17] Office of the National Coordinator for Health IT / NCBI Bookshelf. "Hospital interoperability and exchange data brief." 2024. https://www.ncbi.nlm.nih.gov/books/NBK606033/
[S18] Cruz-Correia R, et al. "Integration of hospital legacy systems through a virtual electronic patient record." BMC Medical Informatics and Decision Making. 2010. https://link.springer.com/article/10.1186/1472-6947-10-15
[S19] Chen Y, et al. "Downtime in digital hospitals." 2017. https://research-management.mq.edu.au/ws/portalfiles/portal/82964539/Publisher_version_open_access_.pdf
[S20] Black Book Market Research. "Improving the patient identification process and interoperability." 2021. https://www.blackbookmarketresearch.com/blog/improving-the-patient-identification-process-and-interoperability-to-decrease-patient-record-error-rates
[S21] Shrank WH, Rogstad TL, Parekh N. "Waste in the US Health Care System: Estimated Costs and Potential for Savings." JAMA. 2019. https://pubmed.ncbi.nlm.nih.gov/31589283/
[S22] Productivity Commission, Australia. "Leveraging digital technology in healthcare." 2024. https://assets.pc.gov.au/research/completed/digital-healthcare/digital-healthcare.pdf
[S23] Abdulsalam Y, Gopalakrishnan M, Maltz A, Schneller E. "Hospital supply expenses: an important ingredient in health services research." Health Care Management Review. https://pubmed.ncbi.nlm.nih.gov/29148349/
[S24] Healthcare Dive. "Hospitals waste billions on bad supply chain management, Navigant says." 2019. https://www.healthcaredive.com/news/hospitals-waste-billions-on-bad-supply-chain-management-navigant-says/567403/
[S25] Overhage JM, McCallie D Jr. "Physician time spent using the electronic health record during outpatient encounters." Annals of Internal Medicine. 2020. https://www.acpjournals.org/doi/10.7326/M18-3684
[S26] Rotenstein LS, et al. "Electronic health record time and visit volume among primary care physicians." JAMA Network Open. 2023. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2812258
[S27] Iscoe MS, et al. "Time spent by emergency physicians using the electronic health record." JAMA Network Open. 2024. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2822299
[S28] Hersh WR, Totten AM, Eden KB, et al. "Health information exchange." JMIR Medical Informatics. 2015. https://medinform.jmir.org/2015/4/e39/
[S29] Agarwal R, et al. "ERP implementation in hospitals: a case study." 2012. https://pubmed.ncbi.nlm.nih.gov/23079029/
[S30] Fiaz M, Ikram A, Ilyas A. "ERP systems: digitization of healthcare service quality." Administrative Sciences. 2018. https://www.mdpi.com/2076-3387/8/3/38
[S31] Engelmann S, et al. "Expanding a major ERP system toward real-time operation planning." 2017. https://pmc.ncbi.nlm.nih.gov/articles/PMC6241744/
[S32] "ERP integration in a healthcare environment: a case study." Journal of Enterprise Information Management. https://www.emerald.com/insight/content/doi/10.1108/17410390610636913/full/html
[S33] Australian Commission on Safety and Quality in Health Care. "Evidence briefing: Electronic medication administration records." 2021. https://www.safetyandquality.gov.au/sites/default/files/resources/attachments/evidence_briefings_on_interventions_to_improve_medication_safety_3_electronic_medication_administration_records_july_2021.pdf
[S34] Australian Commission on Safety and Quality in Health Care. "Evidence briefing: Closed-loop medication management systems." 2021. https://www.safetyandquality.gov.au/sites/default/files/resources/attachments/evidence_briefings_on_interventions_to_improve_medication_safety_4_closed-loop_medication_management_systems_july_2021.pdf
[S35] Poon EG, Keohane CA, Yoon CS, et al. "Effect of bar-code technology on the safety of medication administration." New England Journal of Medicine. 2010. https://www.nejm.org/doi/full/10.1056/NEJMsa0907115
[S36] Jessurun JG, et al. "Effect of automated unit dose dispensing with barcode-assisted medication administration on medication administration errors." International Journal for Quality in Health Care. 2021. https://academic.oup.com/intqhc/article/33/4/mzab142/6400413
[S37] Olson APJ, et al. "Ambient artificial intelligence scribes and clinician documentation burden." 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12492056/
[S38] Tan TF, et al. "Ambient AI scribes and documentation time." JMIR Medical Informatics. 2026. https://medinform.jmir.org/2026/1/e85580
[S39] Rotenstein LS, et al. "Ambient artificial intelligence scribes and clinician EHR time." JAMA Network Open. 2026. https://pubmed.ncbi.nlm.nih.gov/41920565/
[S40] Burdick H, et al. "Effect of a sepsis prediction algorithm on patient outcomes." 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC7245419/
[S41] Calvert JS, et al. "Cost and mortality impact of an algorithm to predict severe sepsis and septic shock." 2017. https://www.tandfonline.com/doi/full/10.1080/13696998.2017.1307203
[S42] Australian Digital Health Agency. "Implementation of AI scribes in healthcare workflows." 2025. https://digitalhealth.org.au/wp-content/uploads/2025/07/Implementation_AI-scribes-in-healthcare-workflows.pdf
[S43] Pelletier D, Duffield C, et al. "Documentation and the transfer of clinical information in two aged care settings." Australian Journal of Advanced Nursing. 2005. https://pubmed.ncbi.nlm.nih.gov/16496835/
[S44] "Reducing and Sustaining Duplicate Medical Record Creation by Usability Testing and System Redesign." Journal of Patient Safety. 2017/2021. https://journals.lww.com/journalpatientsafety/Abstract/2021/10000/Reducing_and_Sustaining_Duplicate_Medical_Record.8.aspx