2026年9月27日 星期日
The Myth of European Autonomy: Why Tony Blair’s Realism Exposes the Fragility of the Old Continent
The British Boarding School Racket: How Elite British Guardianship Turned a Parent's Love into a Fatal Cash Machine
The British Boarding School Racket: How Elite British Guardianship Turned a Parent's Love into a Fatal Cash Machine
英國留學的黑色黑洞:當父母的退休金變成一條冷血的收割流水線
The Price of an Empire: Why South Korea’s Inheritance Tax is a Masterclass in Economic Seppuku
The Price of an Empire: Why South Korea’s Inheritance Tax is a Masterclass in Economic Seppuku
帝國的贖金:為什麼南韓的遺產稅是一場華麗的經濟切腹秀
TOC analysis of GP practice KPIs: Do they improve care or encourage local optimization?
TOC analysis of GP practice KPIs: Do they improve care or encourage local optimization?
The GP KPI framework has the same fundamental TOC risk as the hospital framework: individual measures can improve while the overall patient journey gets worse.
In general practice, the central challenge is to provide safe, effective, continuous care while managing limited clinical capacity, unpredictable demand, and competing needs.
The KPIs cover five important dimensions: safety, effectiveness, caring, responsiveness, and leadership. But they can create conflicts when practices are judged on individual targets without considering how those targets interact.
I will use the KPI descriptions as the basis for the analysis. The examples below identify potential unintended consequences, not evidence that a particular practice is gaming its measures.
1. Local optimization: Where GP KPIs can conflict
The global goal can be expressed as:
Provide patients with timely, safe, effective, compassionate, and continuous primary care, using available resources sustainably.
A. Access targets vs. continuity of care
Potential conflict
More appointments vs. seeing the right clinician
KPIs involved: Appointments within two weeks, same-day urgent appointments, phone-answer times, patient survey scores.
Local optimization: Maximize the number of appointments offered and reduce waiting times.
Potential unintended consequence: Patients may be directed to whichever clinician is available, even when continuity with their usual GP would be beneficial.
Global goal at risk: Coordinated, effective care, particularly for patients with complex or ongoing conditions.
TOC insight: Maximizing appointment throughput does not necessarily maximize the quality or effectiveness of the entire care pathway.
B. QOF targets vs. the patient's overall needs
Potential conflict
Meeting disease-specific targets vs. treating the whole patient
KPIs involved: QOF points, diabetes HbA1c and blood-pressure measures, asthma reviews, screening and immunisation targets.
Local optimization: Focus clinical time on patients and activities that improve QOF achievement.
Potential unintended consequence: Other needs—such as complex symptoms, mental health, multimorbidity, or care coordination—may receive less attention if they are not adequately represented by the targets.
Global goal at risk: Comprehensive, patient-centred care.
TOC insight: A disease-specific indicator is a diagnostic measure, not a complete measure of the patient's overall health.
C. Referral targets vs. appropriate clinical judgment
Potential conflict
Referral activity vs. the right care at the right time
KPIs involved: Cancer referral rates, waiting times, prescribing measures, and clinical outcome indicators.
Local optimization: Focus on meeting referral or prescribing expectations.
Potential unintended consequence: Clinical decisions may be influenced by the measure rather than solely by the patient's needs, if the target is poorly designed.
Global goal at risk: Appropriate, evidence-based care.
TOC insight: The purpose is not to maximize or minimize referrals in isolation. It is to ensure that patients receive the right assessment and treatment without avoidable delay.
D. Phone-answer time vs. clinical capacity
Potential conflict
Answering calls quickly vs. resolving patient needs
KPIs involved: Average phone-answer time, percentage of calls answered, appointments offered versus demand.
Local optimization: Reduce call-answer times by moving callers through the system quickly.
Potential unintended consequence: Patients may still struggle to secure an appropriate appointment or have to call repeatedly.
Global goal at risk: Timely access to effective care.
TOC insight: The time taken to answer the phone is only one part of the patient journey. The more important question is whether the patient gets the care they need.
E. Financial performance vs. clinical capacity
Potential conflict
Controlling expenditure vs. maintaining sufficient capacity
KPIs involved: Finance, staffing, training, appointment availability, and access.
Local optimization: Reduce expenditure to meet a financial target.
Potential unintended consequence: Insufficient staffing or limited appointment capacity can increase waiting times and workload.
Global goal at risk: Sustainable, safe, timely care.
TOC insight: Cost control should support the overall purpose of the practice, not undermine the capacity needed to deliver care.
2. Can GP practices game these KPIs?
Yes. The design of targets can create incentives to improve reported performance without necessarily improving patient outcomes. This can happen through legitimate prioritization, changes in workflows, or potentially inappropriate behavior.
Here are some risks worth examining:
|
KPI
|
Potential gaming or distortion risk
|
| --- | --- |
|
QOF points
|
Concentrating effort on rewarded indicators while less-measured patient needs receive less attention.
|
|
Appointment within two weeks
|
Offering an appointment that meets the target but is not with the appropriate clinician or does not resolve the patient's problem.
|
|
Same-day urgent access
|
Classifying demand in ways that improve reported access without ensuring that all urgent patients receive appropriate care.
|
|
Phone-answer time
|
Shortening calls or redirecting callers without resolving their underlying need.
|
|
Antibiotic prescribing
|
Focusing on a prescribing threshold without adequately accounting for clinical complexity or patient circumstances.
|
|
Patient survey scores
|
Focusing on satisfaction ratings rather than the underlying causes of poor access or continuity.
|
|
Significant events logged
|
Treating the number of recorded incidents as a performance target, rather than focusing on learning and prevention.
|
|
Referral rates
|
Interpreting high or low referral rates as inherently good without considering clinical appropriateness.
|
The key distinction is between improving the metric and improving the outcome the metric is intended to represent.
A low number of reported incidents, for example, is not necessarily evidence of safer care. A practice that encourages staff to report and learn from incidents may initially record more events while improving its safety culture.
3. Uncertainty, conflict, and complexity in general practice
A. Uncertainty: Demand is variable, but access targets are often fixed
GP practices face unpredictable demand from acute illness, chronic disease, seasonal infections, urgent mental health needs, and patients with complex conditions.
A fixed access target does not tell a practice how to manage that variability.
Potential trap: Treating every day as if demand were predictable and trying to maximize appointment utilization.
TOC response: Identify the constraint in the appointment system, understand demand patterns and variability, and protect sufficient capacity to handle urgent and unpredictable needs.
B. Conflict: Clinical priorities vs. performance targets
A GP may need to choose between:
Seeing an urgent patient today.
Reviewing a patient with a chronic condition.
Maintaining continuity for a patient with complex needs.
Completing preventive care and QOF-related work.
All may be legitimate priorities, but they compete for limited clinical time.
Potential trap: Treating each target as equally urgent and expecting staff to meet all of them without resolving the underlying capacity conflict.
TOC response: Establish clear clinical priorities and align appointment allocation with the overall goal of safe, effective care.
C. Complexity: A GP practice is part of a wider healthcare system
A practice's performance depends on services outside its control, including:
Hospital specialist appointments.
Diagnostic services.
Community nursing.
Mental health services.
Social care and support services.
A GP may identify a patient's need and make a referral, yet the patient may still face a long wait for the next stage of care.
Potential trap: Judging the practice only on its own activity or referral measures without considering the complete patient journey.
TOC response: Identify where the patient pathway is constrained, distinguish what the practice can control from what it cannot, and coordinate improvement across the relevant services.
4. The GP access dilemma: An Evaporating Cloud
A simplified TOC Evaporating Cloud helps expose the conflict between rapid access and continuity.
A — COMMON GOAL
Provide timely, safe, effective care that meets patients' needs.
B — NEED 1
Ensure patients can access care promptly, especially when urgent.
C — NEED 2
Ensure patients receive appropriate, continuous, personalized care.
D — ACTION 1
Offer the next available appointment to reduce waiting time.
D′ — ACTION 2
Prioritize continuity with the patient's usual clinician.
THE APPARENT CONFLICT
Rapid access may appear to require sacrificing continuity, while continuity may appear to require accepting longer waits.
The assumption to challenge
The conflict depends on the assumption that every patient must be allocated either to the next available clinician or to their usual GP.
That may not be necessary.
A practice could potentially differentiate between:
Urgent problems that require prompt assessment.
Ongoing conditions where continuity is particularly valuable.
Routine administrative or preventive care.
Complex cases requiring coordinated appointments.
The objective is not to maximize continuity or access in isolation. It is to design an appointment system that delivers both where clinically appropriate.
This is a possible direction for improvement, not a claim that every practice has sufficient capacity to implement it without additional resources.
5. What would a TOC-aligned GP KPI system look like?
Rather than replacing the five CQC questions, I would connect them to a clear hierarchy of measures.
Level 1 — Global goal
Are patients receiving the care they need?
Measure safety, clinical outcomes, timely access, continuity, patient experience, and equity.
Level 2 — Patient flow
Where are patients waiting, and why?
Measure time from requesting care to appropriate assessment and resolution, including delays caused by referrals or other services.
Level 3 — Constraint indicators
What is limiting the practice's ability to deliver care?
Examine clinical staffing, appointment capacity, demand variability, telephone access, administrative workload, and referral bottlenecks.
Level 4 — Guardrails
Are improvements creating unintended harm?
Monitor prescribing safety, clinical outcomes, continuity, patient experience, safeguarding, and financial sustainability.
Five practical changes
Connect access targets to clinical outcomes. Measure whether patients receive appropriate care, not just whether they get an appointment.
Protect continuity where it matters clinically. Distinguish patients who benefit from a familiar clinician from those whose needs can be met by another appropriate professional.
Identify the true constraint. Determine whether the bottleneck is appointment capacity, telephone access, clinical staffing, administrative work, or an external service.
Use QOF as a diagnostic tool, not a complete definition of quality. Ensure that disease-specific targets do not crowd out broader patient needs.
Measure the whole patient journey. Track unresolved needs and downstream delays, not just activity completed within the practice.
6. The most important TOC insight
The GP KPI system risks confusing activity with achievement.
More appointments do not necessarily mean more problems resolved.
Higher QOF scores do not necessarily mean every patient receives comprehensive care.
Faster telephone answering does not necessarily mean better access.
Fewer reported incidents do not necessarily mean safer care.
A balanced budget does not necessarily mean the practice has enough capacity to meet patient needs.
The central question is:
If a GP practice improves this KPI, does the patient's overall health and care experience improve—or does the problem simply move somewhere else?
Bottom line
The five CQC questions provide a useful framework for assessing quality, but the individual KPIs can encourage local optimization if they are treated as independent goals.
A TOC approach would connect them to a shared global goal, identify the constraint limiting patient care, and align appointment allocation, clinical priorities, and resources around improving the entire patient journey.
The Grand Hotel Conundrum: Why Evicting Illegal Migrants to Free the NHS Makes Cynical, Brilliant Sense
The Grand Hotel Conundrum: Why Evicting Illegal Migrants to Free the NHS Makes Cynical, Brilliant Sense
交換房間的黑色喜劇:為什麼把難民請出飯店、讓病人住進去,才是救了英國NHS的唯一解方
The Bed-Blocking Bureaucracy: Why Hospitals Are Just Expensive Hotels for People Who Can't Go Home
The Bed-Blocking Bureaucracy: Why Hospitals Are Just Expensive Hotels for People Who Can't Go Home
佔著茅坑不拉屎的醫療荒謬劇:為什麼醫院成了全天下最貴的養老院
The Phantom Bottleneck: Why Hospitals Squeeze the Wrong Pipes and Wonder Why Everything Floods
The Phantom Bottleneck: Why Hospitals Squeeze the Wrong Pipes and Wonder Why Everything Floods
塞車的醫院與盲目的官僚:為什麼把病床塞滿只會換來全院癱瘓
The 95% Occupancy Illusion: Why Hospitals Worshiping Full Beds Are Actually Breeding Chaos
The 95% Occupancy Illusion: Why Hospitals Worshiping Full Beds Are Actually Breeding Chaos
爆滿的病床幻覺:為什麼醫院追求百分百滿床,只會換來集體崩潰
The KPI Trap: Why Corporate Dashboards Invent Better Excuses Than Results
The KPI Trap: Why Corporate Dashboards Invent Better Excuses Than Results
績效指標的荒謬劇:為什麼公司報表永遠在獎勵說謊?
TOC analysis: Do the CQC KPIs conflict, encourage local optimization, or fail to manage uncertainty, conflict, and complexity?
TOC analysis: Do the CQC KPIs conflict, encourage local optimization, or fail to manage uncertainty, conflict, and complexity?
The KPI framework contains several potential conflicts and risks of local optimization. The central TOC question is whether each measure helps the healthcare system achieve its global goal—or whether it encourages an individual department, hospital, or trust to improve its own numbers at the expense of the whole.
I would examine the indicators through three TOC lenses:
Local optimization vs. the global goal: Do measures reward one part of the system while making overall patient care worse?
The possibility of gaming: Can a hospital improve its reported numbers without improving the underlying outcome?
Uncertainty, conflict, and complexity: Do the KPIs address the root causes of poor performance, or do they try to control symptoms and variability through targets?
One important distinction: the examples below identify potential structural risks, not evidence that a particular NHS trust is actually manipulating its metrics.
1. Local optimization: When KPIs conflict with the global goal
For this analysis, the global goal can be expressed as:
Deliver safe, effective, compassionate, and timely care to patients, using available resources sustainably.
The five CQC questions cover important dimensions of that goal. The risk arises when an individual KPI becomes a target in its own right, rather than a signal of whether the system is achieving the goal.
A. Examples of conflicting KPIs
Conflict 1
A&E waiting time vs. safe, effective care
KPIs involved: A&E 4-hour target, 12-hour trolley waits, patient safety, patient outcomes.
Local optimization: Move patients out of A&E quickly to improve the waiting-time statistic.
Potential unintended consequence: Patients may be transferred to wards before the receiving service is ready, or discharged before they are sufficiently stable.
Global goal at risk: Safe, effective care and continuity of treatment.
TOC insight: Improving the time spent in one department does not necessarily improve the total time from arrival to definitive treatment. The constraint may be downstream, such as a lack of staffed beds.
Conflict 2
Bed occupancy vs. patient flow and resilience
KPIs involved: Bed occupancy, A&E waits, cancellation rates, staffing levels.
Local optimization: Keep beds occupied to maximize apparent utilization of hospital assets.
Potential unintended consequence: Reduced capacity to absorb emergency admissions, delayed transfers, and increased congestion.
Global goal at risk: Reliable, timely care when demand fluctuates.
TOC insight: Maximizing utilization of every resource is not the same as maximizing system throughput. Capacity buffers may be necessary to absorb variation.
Conflict 3
Financial balance vs. safe staffing
KPIs involved: Financial deficit/surplus, nurse vacancy, staffing levels, infection rates, patient safety.
Local optimization: Reduce staffing expenditure to meet a financial target.
Potential unintended consequence: Increased workload, delayed care, staff sickness, and greater safety risks.
Global goal at risk: Safe, effective care delivered sustainably.
TOC insight: A financial target is a constraint on available resources, not the ultimate purpose of the healthcare system. The relevant question is whether spending decisions improve overall system performance.
Conflict 4
Short-term throughput vs. long-term patient outcomes
KPIs involved: Discharge rates, readmission rates, mortality, length of stay.
Local optimization: Discharge patients quickly to free beds.
Potential unintended consequence: If discharge planning or community support is inadequate, patients may return to hospital or experience avoidable harm.
Global goal at risk: Effective recovery and continuity of care.
TOC insight: The system should optimize the entire patient journey, not just the hospital's internal length of stay.
Conflict 5
Meeting a treatment-time target vs. prioritizing clinical need
KPIs involved: Cancer 62-day treatment, referral-to-treatment waits, diagnostic waiting times.
Local optimization: Prioritize cases that are easiest to complete within the reporting period.
Potential unintended consequence: Complex patients or patients requiring multiple diagnostic steps may receive less attention.
Global goal at risk: Timely treatment based on clinical need.
TOC insight: The measure should encourage removal of the constraint in the patient pathway, not merely the completion of cases that are easiest to process.
The common pattern: Each KPI can be reasonable in isolation, yet the combined system can encourage contradictory decisions. A hospital may improve a department's performance while worsening the patient's experience across the entire pathway.
2. Can people game these KPIs?
Yes, the design of performance measures can create opportunities for gaming, particularly when targets are tied to financial rewards, sanctions, reputation, or managerial evaluation.
Gaming does not necessarily mean falsifying data. It can mean changing behavior to improve the measured result without improving the underlying outcome.
|
KPI
|
Potential gaming or distortion risk
|
| --- | --- |
|
A&E 4-hour target
|
Changing patient routing or categorization, or prioritizing the clock over the full care pathway.
|
|
12-hour trolley waits
|
Moving patients to another location or changing how waits are recorded, without resolving the underlying capacity problem.
|
|
Bed occupancy
|
Treating high occupancy as success, even when it leaves insufficient capacity to absorb variation.
|
|
Readmission rate
|
Avoiding appropriate readmissions or shifting care burdens to other services.
|
|
Cancer treatment target
|
Prioritizing cases that can be completed within the target window over more complex cases.
|
|
Financial balance
|
Deferring necessary expenditure or shifting costs to another department or organization.
|
|
Patient satisfaction
|
Focusing on survey scores or selectively collecting feedback rather than addressing underlying patient experience.
|
|
Nurse vacancy
|
Reducing the reported vacancy rate without ensuring sufficient skilled staff are available on each shift.
|
These are risks to investigate, not claims that these behaviors occur at any particular hospital.
The TOC problem with target-driven measurement
A KPI becomes dangerous when it is treated as a goal rather than as a measure of performance.
A more useful distinction is:
Outcome measure: Did the patient receive better care?
System performance measure: Did the entire patient pathway improve?
Diagnostic measure: What is preventing the system from improving?
Compliance measure: Was a required standard met?
If a KPI measures only compliance or activity, it may fail to reveal whether the global goal is being achieved.
3. Are the KPIs falling into the traps of uncertainty, conflict, and complexity?
A. Uncertainty: Treating variability as if it were predictable
Hospitals face fluctuating emergency demand, unpredictable patient acuity, staffing absences, variable lengths of stay, and uncertain discharge needs.
A fixed target such as a waiting-time threshold does not, by itself, explain how to manage those fluctuations.
Potential trap: Treating every breach as a failure of execution rather than distinguishing between:
A recurring capacity constraint.
A temporary demand surge.
A mismatch between staffing and patient needs.
A bottleneck elsewhere in the patient pathway.
TOC response: Identify the constraint, understand the variation affecting it, and protect the flow of patients through the system. Capacity buffers and prioritization rules should be designed around the actual constraint rather than around a blanket assumption of maximum utilization.
B. Conflict: Setting targets that force departments to make incompatible decisions
The hospital may simultaneously be expected to:
Reduce waiting times.
Minimize bed occupancy.
Increase throughput.
Reduce costs.
Maintain safe staffing.
Improve patient outcomes.
These goals can pull managers in different directions.
For example, the emergency department may need to transfer patients quickly, while the receiving ward may be under pressure to maintain safe staffing and avoid exceeding its capacity.
Potential trap: Treating the conflict as a problem of insufficient effort or poor cooperation, when the underlying issue is that both departments are responding rationally to different performance measures.
TOC response: Identify the shared goal and expose the assumptions behind the conflicting actions. Then find a way to satisfy both legitimate needs without sacrificing the global goal.
C. Complexity: Managing a network of interdependent services through isolated KPIs
A hospital is not a collection of independent departments. Emergency care, diagnostics, surgery, wards, discharge services, community care, and primary care are interconnected.
A change in one part of the system can create consequences elsewhere.
For example:
Faster diagnostics can increase demand for specialist consultations.
Faster surgery can increase demand for postoperative beds.
Faster discharge can increase demand for community services.
A staffing shortage in one specialty can delay multiple patient pathways.
Potential trap: Creating a long list of KPIs and assuming that improving each one independently will improve the whole system.
TOC response: Identify the system's current constraint, understand the dependencies around it, and subordinate the rest of the system's decisions to improving overall flow. Reassess the constraint when conditions change.
4. The underlying dilemma: A TOC Evaporating Cloud
The conflicting pressures can be represented by a simplified Evaporating Cloud.
A — COMMON GOAL
Deliver safe, effective, timely care while using resources sustainably.
B — NEED 1
Protect patients and maintain clinical quality.
C — NEED 2
Maintain timely access and sustainable resource use.
D — ACTION 1
Protect clinical capacity and avoid unsafe transfers or discharges.
D′ — ACTION 2
Accelerate patient flow and release capacity for incoming demand.
THE APPARENT CONFLICT
Both actions appear necessary, but if each is pursued independently, one may undermine the other.
The key TOC question
What assumption makes these actions appear mutually exclusive?
One possible assumption is that faster patient flow necessarily means compromising safety, or that protecting safety necessarily means accepting delays.
A potential breakthrough is to challenge that assumption: Can the system remove the constraint that creates the trade-off?
For example, if delayed discharge is caused by a lack of community care capacity, increasing pressure on the emergency department or ward may not solve the underlying problem. Improving the discharge pathway could support both safety and flow.
This is a hypothesis to test against actual patient-flow data, not a conclusion about any particular trust.
5. How I would redesign the KPI system using TOC
Rather than abolishing the five CQC questions, I would distinguish global outcome measures, constraint measures, and diagnostic measures.
A TOC-aligned measurement framework
Level 1 — Global goal
Are patients receiving safe, effective, timely care?
Measure outcomes across the entire patient journey, including avoidable harm, recovery, access, and continuity of care.
Level 2 — System flow
Where is the constraint preventing better outcomes?
Measure total patient journey time, delays at bottlenecks, flow interruptions, and the capacity of the constrained resource.
Level 3 — Diagnostic indicators
What is causing the constraint or limiting performance?
Use staffing, bed occupancy, diagnostic capacity, discharge delays, infection rates, and other indicators to explain the system's performance.
Level 4 — Guardrails
Are improvements creating unacceptable side effects?
Monitor safety, clinical quality, patient experience, equity, and financial sustainability so that improving flow does not undermine other essential needs.
Five practical changes
Stop treating every KPI as an independent target. Make the global goal explicit and show how each measure contributes to it.
Identify the constraint before deciding where to intervene. The bottleneck may be outside the department with the worst reported performance.
Measure end-to-end patient flow. Track the patient's complete journey rather than relying solely on department-level clocks.
Use KPIs to diagnose causes, not just assign blame. Distinguish capacity shortages, demand variability, process problems, and avoidable delays.
Review whether the constraint has moved. An intervention that improves one bottleneck may reveal another, requiring a new round of improvement.
Bottom line
The central TOC criticism is not that the CQC measures the wrong things. Safety, effectiveness, caring, responsiveness, and leadership are all important.
The risk is that a collection of individually reasonable KPIs can create a system of conflicting incentives, local optimization, and unintended consequences.
A TOC-based approach would preserve the essential quality standards while connecting them to a clear global goal, identifying the system's constraint, and using diagnostic measures to improve the whole patient pathway.
The key question for each KPI is:
If this number improves, does the patient's overall outcome improve—or could the improvement simply shift the problem somewhere else?
That is the distinction between managing the performance of individual parts and managing the performance of the whole system.
The Fraternity Trap: Why Elite Clubhouses Invented Ancient Tribal Grooming for Modern Youth
The Fraternity Trap: Why Elite Clubhouses Invented Ancient Tribal Grooming for Modern Youth
大學兄弟會的原始迷陣:當高級俱樂部變成現代部落的狼性溫床
The Confession Trap: How the FBI and the Ballot Box Killed the Overseas Godfather
The Confession Trap: How the FBI and the Ballot Box Killed the Overseas Godfather
坦白的陷阱:一場心理戰如何讓海外老大哥走入歷史
When the Old Godfather Stopped Working: How Washington Ignored the Republic of China and Broke Overseas Chinese Politics