Quality Intelligence Workspace
Quality intelligence workspace for reporting, approval journey analysis, sampling, and privacy-safe calibration.
Data governance mode
Imported operational data may be saved in this browser for recovery until you close or clear the dataset. Basic master data and anonymized calibration patterns persist until intentionally deleted.
Executive overview
Current period
Load an operational file to see the current quality picture.
Current operational priorities
Ranked signals requiring investigation in the selected period.
View supporting case population
Data, imports and advanced controls
Client complaint intelligence
Import complaint data
Review detected columns, confirm the mapping, then build the analysis.
1Select file2Confirm mapping3View analysis
Choose a complaint workbook. Review the detected mapping and confirm import. Files stay in this browser.
Import locally; every Excel sheet is read in one pass. Case Source classifies each row: IA Portal is Authority, other named channels are Care, and blank source remains Unclassified. Case ID and date are required. Unknown group, department and physician remain visible rather than inferred.
Choose a file to review its columns.
Check the detected columns above before importing.
Shared QInsight data
Import practitioner requests
Independent request table for provider, physician and complaint analysis.
Review the detected approval, provider and practitioner columns. Confirm to save the request extract locally.
Approval, provider and practitioner IDs are required for a verified join.
Approve the provider, approval and practitioner columns shown above.
Practitioner request data No requests imported
No practitioner requests imported. Use the import control above.
| Approval | Provider ID | Practitioner ID | Practitioner name | Qualification | Created |
|---|
Operational detail
Filtered approval population
0
Unique approvals in scope
Repeated-response population
0%
Cases with more than one response
Decision-transition cases
0
Rejected then positive outcome
NPR numerator / denominator
0%
No-preauthorization-required ratio
Estimated data completeness
—
Mapping and event coverage estimate
Service-change impact and evidence integrity
Tracks service composition, requested-service changes across separate submissions, and follow-up intensity. Only approvals with dated service records are included.
Supporting evidence for current priorities
Supporting interpretation
Top response reasons
Highest-frequency approvals
Click a row to open the request journey.
| Reference | Setting | Provider | Policy | Main processor | Responses | Status |
|---|
Case Intelligence
One clear view of the decision, evidence, requested services and responsible processing touches.
Case summary
Assessment
Approval activity journey
First recorded event through the latest submission or decision. Not patient waiting time or TAT.
SubmissionFollow-upOutcome
Request story
Autoplays as a continuous request story. Hover or tap a point for the full response.
Requested services
Processor response touches
Processor Performance Review
Processor-attributable activity, query outcomes, recurring patterns and team comparison.
Core measures
Defect focus
Actionable processor signals only. NPR submission volume is assessed in Provider analysis.
Further-details actions
| Reference | Provider | Service/topic | Processor actions | Total responses | Processor response | Later outcome | Actions |
|---|
Top 25 processor drill-down appendix
| Reference | Provider | Service/topic | Processor actions | Total responses | Response category | Later outcome |
|---|
Executive performance review
Evaluation trend
Response intensity trend
Selected processor compared with the configured team peer average over time.
Measures average recorded responses per approval over time for the selected processor and the configured team peers. Daily, weekly, or monthly values are total responses divided by approvals in that period — a workflow-intensity measure, not call handling time or patient waiting time.
Service / Topic Decision-Change Intelligence
Compared with processors in the same configured team. Signals require at least three queried approvals.
| Service/topic | Queried approvals | Providers | Selected rate | Peer rate | Later positive | Interpretation |
|---|
Top 15 processor service-pattern appendix
| Service/topic | Queried | Providers | Selected rate | Peer rate | Later positive |
|---|
Management interpretation
Top 10 further-detail comments
The most frequently repeated verbatim further-detail comments issued by this processor. Useful for spotting recurring canned responses or drafting a standard clarification template.
Manual QA review evidence
Provider and Provider-Group Report
Analyze an individual provider or a persistent provider group.
Internal diagnostic view · Pareto · Histogram · Run · Control · drill-down
Four PDF configurations available. Uses current provider and period filters.
Top 10 providers requiring immediate behavior review
Providers below 5 approvals are hidden; 5–14 approvals may appear only as emerging low-volume evidence; 15+ approvals are eligible for priority ranking; 30+ may receive high-severity wording.
| Priority | Provider | Review score | Follow-up | Transitions | Why prioritized |
|---|
Frequency intelligence
Transaction frequency and manual processor contribution use unique approval denominators.
Frequency rate over time
Selected processors' share of manual transactions
Submission and declaration intelligence
Phrase-based operational measures. Exact matched comments remain visible for audit.
Top declaration-question processors
Top 20 declaration comments
What is driving the signal?
Manual review evidence
Manual review findings
Provider learning curve
Period snapshots are retained locally by provider ID. Operational trends are shown separately from manual-review findings so the denominator remains clear.
Provider intervention marker
Recommended behavior-review cases
Inpatient activity is adjusted for expected daily authorization follow-up. Reviewed references are removed from future recommendations.
Excess or delayed authorizationAbove two same-day responses or delayed initial positivity
Multiple negative to positiveRepeated negative responses before approval
Positive to negativePossible processor or service-change signal
Provider benchmark comparison
Top contributors
Processors linked to the selected provider’s highest-risk behavior.
Policies carrying the strongest provider-specific behavior signal.
Top involved services by risk factor
Service frequency is separated from NPR and decision-change signals to keep the interpretation clear.
Policy 360 — Utilization, Cost & Risk Intelligence
Policy-level authorization intelligence across members, providers, services, diagnoses, approval performance, frequency, inpatient/OPD utilization and forward risk. Projections are authorization-based and are not actuarial claims forecasts.
Utilization and financial trends
Smooth curves use the selected reporting interval. Point labels and hover values reflect the exact underlying data.
Top utilizing members
Member identifiers are masked in this management view.
Provider consumption
Top OPD services
Top inpatient services
Highest-cost diagnoses
Approved exposure is linked to approvals carrying each diagnosis; diagnosis totals should not be summed across rows.
Authorization performance
Reasons and documentation drivers
Frequency intelligence
Risk intelligence and projection
Interpretation guardrails
Utilizing members are members present in authorization data. A true policy utilization rate requires an eligible-member census denominator. Cost projections are based on authorization history in the selected scope and should not be interpreted as actuarial claims forecasts. Diagnosis exposure is linkage intelligence and may overlap when one approval carries multiple diagnoses.
Inpatient Authorization Intelligence
Reference-level authorization signals. Expected daily inpatient follow-up is excluded from alarms; inferred authorization span is not actual length of stay.
What counts as an alarm?No positive result within the first two responses, more than two responses on one day, repeated negatives before positivity, or a later negative after a positive response.
What is excluded?Routine daily follow-up with timely positive authorization remains visible as workload but does not count as a defect signal.
What should happen next?Use the references below for focused manual review before attributing the signal to a hospital or processor.
Selected-hospital signal profile
How the active hospital scope contributes to inpatient authorization alarms.
Processors linked to the strongest signals
Contribution means involvement in the journey, not confirmed processor responsibility.
Final disposition: home health care
Authorization references whose final recorded response explicitly indicates discharge, transfer, or referral to a home health care setting. This is a response-text signal, not a complete discharge census.
Priority references requiring further review
High-confidence review candidates are ranked by severity. Critical-severity cases remain eligible when timeline evidence is incomplete. Confirm or retire each candidate; retired rows are replaced from the top-50 pool.
Evidence interpretation
High-Cost Inpatient Intelligence
Predicted authorized admission days from final-response stay services, elective PKG classification, requested/approved financial exposure and provider trends. Predicted days are authorization intelligence, not actual length of stay.
Elective PKG classification
Enter exact elective PKG service codes separated by commas. QInsight also recognizes codes explicitly containing the PKG token. Classification is based on service-code evidence, not free-text assumptions.
Admission-day and financial trends
Provider / hospital comparison
High-cost inpatient approval population
One row per unique approval. Predicted days prioritize qualifying admission/ward/ICU/per-diem service evidence and use the response timeline as corroboration.
Methodology
Predicted admission days use only qualifying inpatient stay services in the final response/service snapshot. Earlier submissions and response-timeline span are excluded from the day calculation. Where final-response stay services contain quantity evidence, approved quantity is primary and requested quantity is used only when approved quantity is unavailable. The largest qualifying stay-service quantity is used to avoid double-counting overlapping ICU/ward/room services; otherwise unique qualifying stay services are counted as fallback evidence. Elective PKG classification is code-driven. Total requested and total approved are calculated across the filtered unique approvals.
NPR Hospital Analysis
Analyze all providers, one specific provider, or a saved provider group within a selected date range.
NPR approvals in covered period
0
NPR volume
NPR ratio in covered period
0%
NPR / covered approvals
Covered approvals
0
Denominator
Highest period
—
Peak ratio
Trend direction
—
Slope across periods
NPR ratio over time
Provider NPR comparison
Each hospital is represented horizontally and colored according to its NPR percentage.
Provider NPR timelines
One line per included hospital or branch in the selected provider scope.
Advanced filters, thresholds and data reconciliation
NPR color thresholds
Below 5% — dark green
5% to <7% — pale green
7% to <10% — yellow
10% or above — red
Service-category filter
Selecting a category updates the NPR ratio, trend, provider comparison, policies, services and interpretation using unique approvals.
Contribution to total NPR by service category
Category contribution trend
Each point shows the category share of NPR approvals in that day, week, or month.
Highest-involved policies
Highest-involved services
NPR Summary
Chart-ready data
| Period | Total approvals | NPR count | NPR ratio | Actions |
|---|
Manual Review Center
Generate a representative sample by processor or provider, add extra references manually, and retain reviewed selections in the reporting export.
Report setup
Reviewer not entered
Sample filters
Add request
Selected approval
Guided case review
Review the case, identify involved subjects, then complete focused assessments.
Case verdict
Subjects to assess
Select each processor or provider requiring a focused assessment.
Next: complete the selected subject assessments.
Subject assessment
Assess one subject at a time. Completed subject reviews collapse automatically into their tab.
Recommended knowledge review
ACA-backed guidance is a learning support for processors, not a disciplinary finding. A reviewer must confirm every recommendation.
Presentation evidence example
Only reviewer-approved examples are used in certified presentations. Patient-identifying information and unrestricted notes are excluded.
Case-specific communication
This case-specific message is retained only in the current review/report export. It is not stored in the anonymized calibration store.
Approval reference, patient/member ID, provider ID, processor ID, exact timestamp, and free-text notes are never retained in the calibration store.
Save to automatically open the next pending case.
Review workflow
Pending cases remain compact. Completed reviews show a two-line preview and can be edited.
0 reviewed
Pending cases
Completed reviews
Reviewed case registry
Completed reviews and imported historical quality-signal cases remain searchable across processor/provider changes. Signal-history cards preserve prior evidence without fabricating missing review fields.
Manual Review Feedback Report
In favorAgainstMixed
Insights and Calibration
Investigate clarification effectiveness, then retain reviewer-confirmed learning in the privacy-safe calibration store.
Clarification Effectiveness and Avoidable Follow-up
Fixed statements are evaluated against the next recorded outcome. Rates use unique approvals with at least one clarification.
Outcome trend
Weekly share of clarified approvals by their first later outcome.
Top processors
Top providers
Top services
Exact evidence cases
Retained patterns
0
Privacy-safe reviews
Root-cause agreement
—
No evaluable reviews
Avoidability agreement
—
No evaluable reviews
Unresolved
0
Require further review
New since export
0
Incremental package
Agreement by system prediction
Calibration status
Coverage readiness
Privacy validation
View reviewer consistency signals
Reviewer-consistency monitoring contains identifiable reviewer IDs, remains local to this browser profile, and is not role-based access control. It is not included in privacy-safe calibration exports.
Prediction accuracy
Prediction accuracy
Shows how often system conclusions agree with manual review. Non-evaluable cases are shown separately and do not affect the percentage.
Root-cause agreement
—
No evaluable reviews
Avoidability agreement
—
No evaluable reviews
Retained anonymized patterns
Retained anonymized patterns
No private identifiers or raw notes are stored.
| Period | Setting | Processor team | Service band | Prediction | Reviewer conclusion | Confidence |
|---|
Basic Application Setup
Processor teams and provider groups persist across normal review resets.
Report Designer
One layout path from live report → canvas → PDF. The selected template is also the print template; there is no separate assignment field.
System Default
Report Designer
One layout path from live report → canvas → PDF. The selected template is also the print template; there is no separate assignment field.
Actual A4 portrait page
210 × 297 mm · 6 mm printable safety margin · free mouse placement
Page 1 / 1
Import processor masterExcel / CSV
Excel / CSV processor list
Expected: Team Name, Manager ID, Manager Name, Region Code, Member User ID, Member Name, Email, Role Code/Role, Active Status, Eligible Reviewer, Exclude From Executive Analytics, Effective From, Effective To.
Import provider masterExcel / CSV
Excel / CSV provider list
Map the provider ID, provider name, group, and optional region columns after selecting the file.
Detected provider-list mapping
Import policy masterExcel / CSV
Excel / CSV policy list
Map the policy number and policy name. Imported names are applied to matching approval data and exports.
Detected policy-list mapping
Processor organizational master
Import the organizational master or edit one processor at a time. Member User ID is the stable processor key.
CX Assist processor aliases
Map a CX Assist Agent Name or Agent ID to the matching QInsight processor. Aliases are kept locally and applied to future imports.
Executive analytics exclusions
Provider groups
Search the saved provider list, select providers, then assign or remove a group without re-importing data.
0 providers
ProviderGroupRegion
Operational KPI progression
Set the current achievable KPI thresholds used in processor reports. Ultimate operational targets remain visible as the long-term destination.
Ultimate target ≤ 6.0%
Ultimate target ≤ 9.0%
Ultimate target ≤ 1.60
Ultimate target ≤ 1.90
Ultimate target ≤ 1.75
Report visual assets
SVG
Automatic uses the dark certified logo on light report headers and the white certified logo on dark report surfaces.
Data lifecycle controls
Close & clear current operational dataset
Deletes loaded operational data while preserving masters and calibration knowledge.
Clear current file without learning
Use when the file was incorrect or only tested.
Full application reset
Deletes operational data, masters, groups, and calibration store.
Client Complaints
Trace complaints from hospital to physician, service and decision; review the supporting cases.
Complaint measure reconciliation
What the selected complaints show
First recorded request to complaint
Elapsed hours from the first dated Time Track event to complaint Created On, for uniquely linked approvals with exact timestamps. Fixed review levels: 24 h lower, 48 h target, 72 h upper.
Complaint channels and validation
Two views of the same cases. Select a slice to filter; open the table for exact counts.
Departments and practitioner specialties · complaint contribution
Qualification Display on a uniquely linked practitioner request determines department. An explicit reported department is used only when qualification is unavailable; service values never become departments. Select a bar for its cases.
Services involved in complaints
Compare all requested services with services on complaint-linked approvals. One approval may list several services.
Top 10 physicians in this selection
Select a hospital or group to rank its physicians. Request rates require a unique practitioner link.
Service category mix
Current category cards retain the established QInsight method while coded complaint targets are checked in shadow mode. They describe recorded cases, not verified target-service shares. Unclassified cases stay in the denominator.
Top 15 specific services
Top service–reason combinations
Specific services paired with recorded rejection reasons. Bars show share of all cases; select a row for the exact records.
Complaint volume and cumulative share
Case-count columns and cumulative share with an 80% guide. Select a column for its cases.
Service × documented root cause
See where documented root causes concentrate by service. Select a pair for its cases.
Missing root-cause fields remain outside this cross-link but inside all complaint totals. A count is not a provider-quality rate or proof of cause.
Group → hospital → department / specialty → specific service
Highest-volume paths through the selected hospitals. Select a path for its cases.
Run chart · complaint outcomes
Choose an entity above, then select a period to filter the analysis.
Repeat complaints per approval · histogram
Select a repeat-count bar to filter the analysis; select it again to clear. Cases without an approval remain visible in the unfiltered scope.
Practitioner request context
Affected requests ÷ distinct loaded practitioner requests. Hospital contribution is shown separately.
Repeat-patient pathways across approvals
Patients with complaints on two or more approvals. Select a cluster for its records; identifiers stay hidden.
Processor and approval-reference frequency
Prior processor and approval-reference links. Select a bar for its exact cases; links do not establish fault.
Complaint export context
Internal decision linkage · coverage details
Only a unique provider + approval match with one identified processor on dated pre-complaint events is linked. Shared, same-day and unmatched decisions stay separate; counts do not establish fault.
Shared improvement pathway
A data-led discussion: inspect the case evidence, agree an action and owner, then compare the same measure over an equal follow-up period.
Open full report preview
Complaint records
Select a bar or gap to inspect its records.
Definitions and interpretation
One distinct record is one categorized source plus one case number. Each imported row is classified by Case Source: IA Portal is Insurance Authority; other supplied named channels are Tawuniya Care; missing Case Source remains Unclassified. All three categories are included in overall totals and their shares use all selected complaints as denominator. In Customer Favor is valid, In Tawuniya Favor is non-valid; remaining values are other. Every Pareto denominator includes unattributed cases. Weekly or monthly trend uses Created On and the selected trend interval. Physician charts use a supplied physician field or a unique provider + approval match in the practitioner request extract; ambiguous and unmatched cases remain unattributed. Department contribution uses the unique linked practitioner's Qualification Display first. An explicit reported Department is used only when qualification is unavailable; service values never become department names. Qualification Display names a specialty and is not proof of a hospital's organizational department. Common title words and narrow spelling aliases are normalized for grouping; clinically distinct specialties remain separate and the original workbook value stays in the practitioner table. Cases with neither source remain Unattributed. Conflicting qualifications are excluded from specialty linkage. Specialty request rates count distinct loaded provider + approval + practitioner requests in that specialty and distinct complaint-affected requests; they describe only the loaded extract. A practitioner complaint percentage counts distinct matched requests with at least one complaint divided by that practitioner's distinct loaded requests in the selected request period; complaint case counts remain separate. A doctor's contribution divides their linked complaint cases by all hospital complaint cases in scope. Neither metric establishes fault or clinical quality. A linked-cases-per-100-approvals figure counts only complaint cases whose provider + approval key is present in its loaded practitioner-request denominator; the displayed complaint-case count still includes unmatched cases. QInsight seven-category visuals count each complaint once; specific services on a linked approval are supporting associations and may overlap. The service rankings use total case volume, while distinct affected-approval rates count each matched approval once.
The corrected complaint count is a separate conservative episode view; it never replaces the recorded complaint count or its approved KPIs. It groups cases only when a valid patient ID, same provider or hospital, same attributed physician, same single actual service on a uniquely linked operational approval, and complaint dates within 14 days of the episode's first case are available. Different approval numbers may belong to one episode. Missing identity or physician link stays as a separate case. Unique patient cards require complete member codes in the selected practitioner requests; affected-patient cards require valid IDs on every complaint in that scope. Complaints / approvals is linked complaint cases divided by distinct selected approvals, so repeated cases on one approval can exceed 100%.
Methodology & Calculation Guide
Page-by-page definitions for measures, charts, operational logic, interpretation, and practical use.
High-confidence reviews: 0
Last export: Never