Co Authors

Dr Alwin Tan, GAICD, MBBS, FRACS, EMBA (Melbourne Business School)
Senior Surgeon | Governance Leader | HealthTech Co-founder | Founder of Institute for Systems Integrity (ISI) | Harvard Medical School — AI in Healthcare | University of Oxford — Sustainable Enterprise | Bastas Academy for Healthcare Leadership — Triple Scholar

Nadra Gadeed BSc Psychology, MSc Forensic Psychology, MA Health Research
Founder and Strategic Advisor, Edify & Grow | Workforce Risk Advisor to Boards and C-Suite Leaders, Private Healthcare, GCC & Asia-Pacific | 20 Years in the NHS Across Strategic, Operational and Leadership Roles

INSTITUTE FOR SYSTEMS INTEGRITY

When the People Holding Everyone Else Accountable Become Accountable to Nobody

We investigate doctors.

We investigate nurses.

We investigate executives.

We investigate employees.

We investigate whistleblowers.

We investigate bullying.

Misconduct.

Patient harm.

Professional behaviour.

Fraud.

Safety failures.

Good.

We should.

But there is one group organisations are often much less comfortable investigating:

🚨 THE PEOPLE DOING THE INVESTIGATING.

Who checks their assumptions?

Who challenges their conflicts?

Who asks why one witness was believed and another was not?

Who examines what evidence never made it into the report?

Who asks whether the terms of reference quietly determined the answer before the first interview occurred?

And who distinguishes five genuinely independent complaints from one allegation repeated through five institutional processes?

Credibility judgements are far less reliable than many investigative processes assume. Decades of research into deception detection suggest people distinguish truth from lies at only slightly better than chance, while professional experience or training produces surprisingly limited improvement (Bond and DePaulo, 2006).

An investigation that never asks why one account was considered more believable than another may therefore be relying on human intuition where tested reasoning is required.

And that brings us to the uncomfortable part.

When an investigation is wrong:

🚨 WHO HAS THE POWER TO SAY SO?

Because there is a dangerous moment in every accountability system.

It happens when an institution becomes so confident in its own processes that:

THE PROCESS ITSELF BECOMES PROOF THAT THE PROCESS MUST BE RIGHT.

“We followed policy.”

“The matter was independently investigated.”

“Multiple concerns were raised.”

“Legal reviewed it.”

“The committee considered it.”

“The regulator was notified.”

Every sentence sounds reassuring.

Until someone asks:

DID ANYONE INDEPENDENTLY TEST WHETHER THE UNDERLYING CONCLUSION WAS ACTUALLY CORRECT?

🚨 GOOD PEOPLE CAN RUN BAD PROCESSES

This paper is not an argument against investigators.

Quite the opposite.

Investigations are essential.

External investigators can reduce conflicts.

Independent review can protect complainants, respondents and organisations.

But independent does not mean infallible.

The investigator is still human.

Someone still decides:

who gets interviewed;

what gets asked;

which evidence matters;

which contradictions matter;

whose account appears credible;

what falls outside scope;

and which explanation best fits the evidence.

Research into cognitive bias in workplace investigations suggests professional judgement remains vulnerable to cognitive influences even where investigators act honestly and professionally (MacLean, 2022).

That distinction matters.

The problem is not necessarily corrupt investigators.

The harder governance problem is:

🚨 GOOD PEOPLE CAN RUN BAD PROCESSES.

And:

🚨 HONEST INVESTIGATORS CAN REACH WRONG CONCLUSIONS.

Healthcare already understands this principle.

We do not assume an experienced surgeon is incapable of error.

That is why healthcare developed:

second opinions;

multidisciplinary meetings;

audit;

peer review;

checklists;

morbidity and mortality meetings;

clinical governance.

We do not say:

“The surgeon is experienced. Therefore the surgeon cannot be wrong.”

So why would an organisation effectively say:

“The investigator is independent. Therefore the investigation must be right”?

Expertise deserves respect.

It does not deserve immunity from scrutiny.

🚨 FIVE PROCESSES CAN STILL BE ONE STORY

Imagine this.

One allegation is made.

It enters HR.

HR informs performance management.

Performance management informs credentialing.

Credentialing contributes to a regulatory notification.

The regulatory notification is subsequently referenced as evidence that concerns about the individual were serious.

Now stand back.

There are multiple processes.

Multiple documents.

Multiple committees.

Multiple people discussing the individual.

From the boardroom, it can look compelling.

But follow the evidence backwards.

Where did it actually begin?

Perhaps with one allegation.

This creates what we propose calling:

INSTITUTIONAL PSEUDO-CORROBORATION

The appearance of independent corroboration when multiple organisational processes derive substantially from the same underlying allegation, information or assumption.

Put more simply:

🚨 COPYING AN ALLEGATION INTO FIVE FILES DOES NOT CREATE FIVE PIECES OF EVIDENCE.

Yet systems can accidentally behave as though it does.

Process multiplication becomes mistaken for evidence multiplication.

Repetition becomes mistaken for corroboration.

Institutional activity becomes mistaken for institutional certainty.

Eventually someone says:

“There must be something to this. Look how many processes are involved.”

That should concern anyone responsible for governance.

Because:

🚨 AN INVESTIGATION MUST NEVER BECOME EVIDENCE OF ITS OWN CORRECTNESS.

🚨 WHO WROTE THE QUESTION?

Before an investigator interviews anyone, something extraordinarily powerful may already have happened.

Someone wrote the terms of reference.

Someone defined the allegation.

Someone decided what was inside scope.

Someone decided what was outside scope.

Someone selected the investigator.

Someone supplied the background material.

Someone framed the problem.

Which means:

THE MOST POWERFUL PERSON IN AN INVESTIGATION MAY NOT BE THE INVESTIGATOR.

It may be the person who decided what the investigator was allowed to investigate.

A technically excellent investigation can produce a deeply misleading answer if it was asked the wrong question.

Forensic psychology has long examined this vulnerability. Ask and Granhag (2005) found that investigators experiencing a high need for cognitive closure can become less sensitive to alternative explanations once an initial hypothesis is established.

Bad intent is not required.

Framing alone can matter.

Boards should therefore ask:

Who commissioned the investigation?

Who framed the allegations?

Who selected the investigator?

What information did the investigator receive before interviewing anyone?

What conflicts were identified?

Could the scope change when contradictory evidence emerged?

And perhaps most importantly:

🚨 WHO DECIDED WHAT NOT TO INVESTIGATE?

🚨 AND WHAT IF THE BOARD IS PART OF THE STORY?

There is an even harder version of this problem.

A board commissions an investigation into one of its own members.

Or into an executive it appointed.

Or into conduct arising from a strategy it approved.

Or into a decision it previously endorsed.

Now the organisation's highest accountability body may itself have an interest in the answer.

Independence cannot simply be assumed because the board sits above management.

It must be demonstrated.

That may require external commissioning.

Decision separation.

Explicit conflict management.

Independent assurance.

And, critically, a willingness to accept findings that challenge the board's own previous judgement.

Governance becomes meaningful precisely when accountability can travel upward as well as downward.

🚨 WHAT WOULD CHANGE YOUR MIND?

Perhaps every investigator, committee and decision-maker should be required to answer one question before making a high-consequence finding:

WHAT EVIDENCE WOULD MAKE YOU CONCLUDE THAT YOU ARE WRONG?

Think about that.

If the answer is:

“Nothing.”

Then we may no longer have an investigation.

We have a conclusion looking for confirmation.

Because genuine investigation requires uncertainty.

Evidence must retain the ability to change the answer.

New testimony must matter.

Contradictory documents must matter.

Alternative explanations must matter.

Otherwise the interviews may continue.

The reports may continue.

The committees may continue.

The meetings may continue.

But inquiry has stopped.

🚨 WHEN NOTHING CAN CHANGE THE CONCLUSION, THE INVESTIGATION MAY ALREADY BE OVER — EVEN IF THE PROCESS IS STILL RUNNING.

🚨 BUT HERE IS THE MIRROR WARNING

This argument can itself be weaponised.

A powerful executive accused of bullying can shout:

“Biased investigation.”

A clinician facing legitimate concerns can claim:

“Retaliation.”

Someone accused of misconduct can bury a complainant beneath endless procedural challenges.

That is unacceptable too.

So let us be equally clear:

QUESTIONING AN INVESTIGATION DOES NOT PROVE THE ALLEGATION FALSE.

A flawed process does not automatically mean an innocent respondent.

An investigation does not automatically mean a guilty respondent.

A complaint deserves to be heard.

A respondent deserves to be heard.

A whistleblower deserves protection.

An investigator deserves independence from inappropriate interference.

INTEGRITY REQUIRES ALL FOUR.

Procedural fairness should never become a weapon powerful respondents can use to exhaust genuine complainants.

But neither should the moral importance of protecting complainants make investigative systems immune from scrutiny.

Accountability without fairness fails.

But fairness without accountability fails too.

🚨 WHEN ACCOUNTABILITY ITSELF BECOMES POWER

What if nobody shouts?

Nobody threatens.

Nobody sends an abusive email.

Instead:

a complaint appears.

Then another process.

Then performance management.

Then credentialing.

Then “professionalism concerns.”

Then regulatory reporting.

Every individual step may appear legitimate.

And sometimes every step is legitimate.

But institutional power becomes dangerous when legitimate mechanisms can be repeated, coordinated or inherited without sufficient independent challenge.

Research examining bullying in healthcare has identified fear of being regarded as a troublemaker and beliefs that reporting will not change anything among barriers to speaking up (Carter et al., 2013).

Psychological safety therefore asks:

“Can I speak?”

Organisational safety asks:

“What happens to me after I speak?”

Investigation integrity requires another question:

🚨 WHO CHECKS WHAT THE ORGANISATION DOES NEXT?

🚨 THE ACCOUNTABILITY ASYMMETRY

Consider what happens when an individual is investigated.

The institution may demand:

their emails;

their documents;

their recollections;

their decisions;

their relationships;

their motives;

their explanations;

their credibility.

Now reverse the lens.

Can the individual ask:

Why was this investigator selected?

What conflicts were considered?

What evidence contradicted the allegation?

What alternative explanations were tested?

Why was particular evidence excluded?

Who changed the scope?

Who reviewed the investigator's reasoning?

Who checked whether apparently separate complaints had the same source?

What would have changed the conclusion?

Sometimes the visibility is dramatically different.

We call this:

THE ACCOUNTABILITY ASYMMETRY

The institution demands extraordinary transparency from the individual while providing comparatively little transparency about the machinery judging them.

That is not merely a fairness problem.

It can become a workforce risk.

People watch what happens to colleagues who enter an investigation.

Complainants watch.

Respondents watch.

Whistleblowers watch.

Managers watch.

Future witnesses watch.

And they learn something about what accountability actually means inside that organisation.

Which gives us a governance principle:

🚨 THE MORE ACCOUNTABILITY YOU DEMAND FROM OTHERS, THE MORE ACCOUNTABILITY YOUR OWN SYSTEM MUST BE WILLING TO ACCEPT.

Otherwise accountability flows only downward.

And accountability that flows only downward has another name:

POWER.

🚨 SO WHO INVESTIGATES THE INVESTIGATORS?

Not another investigator every time.

That would risk endless appeals.

Enormous costs.

Delayed resolution.

And further harm to complainants and respondents.

The answer is more sophisticated:

GOVERN THE INVESTIGATION SYSTEM.

Do not simply count investigations.

Assure their integrity.

For high-consequence matters, boards should expect evidence of:

Independence 

— who selected the investigator and what conflicts were considered?

Scope integrity 

— who framed the question and what was excluded?

Evidence provenance

 — are apparently separate concerns genuinely independent?

Bias safeguards 

— what alternative explanations were actively tested?

Procedural fairness 

— did the respondent meaningfully know and answer the case?

Complainant protection 

— were retaliation and repeated harm actively prevented?

Decision separation 

— did too much investigative and decision-making authority sit with one person or group?

Review triggers 

— when does consequence, conflict, uncertainty or new evidence require additional assurance?

This is not about making every investigation longer.

It is about making consequential investigations trustworthy.

🚨 STOP SHOWING BOARDS THE WRONG NUMBERS

Boards commonly receive:

37 complaints.

24 closed.

8 substantiated.

5 ongoing.

Average investigation: 64 days.

Everything green.

Everything measurable.

Everything reassuring.

But those numbers mostly tell the board how efficiently the investigation machinery is moving.

They do not necessarily tell the board whether it is producing reliable outcomes.

The board should also ask:

How many investigations involved identified conflicts?

How many findings changed after independent review?

How many apparently separate complaints originated from common evidence?

How often were alternative hypotheses documented and tested?

How many complainants reported retaliation after speaking up?

How many investigations expanded beyond their original scope?

How many high-consequence findings received independent challenge?

How many complainants or respondents left the organisation within twelve months of an investigation closing?

And above all:

🚨 HOW DO WE KNOW OUR INVESTIGATIONS DESERVE TO BE TRUSTED?

A dashboard can tell you how efficiently your investigation machinery operates.

It cannot automatically tell you whether the machinery is just.

🚨 THE HARDEST THREE WORDS IN GOVERNANCE

There is one final test of institutional integrity.

Can the organisation say:

“WE WERE WRONG.”

Not:

“The process was followed.”

Not:

“The investigator was independent.”

Not:

“The committee considered the matter.”

Not:

“Legal reviewed it.”

Not:

“We acted on the information available.”

But:

WE GOT IT WRONG.

And then correct it.

Because an organisation capable of investigating everyone except itself has not built a mature accountability system.

It has built an authority system.

The strongest accountability system is not one that never makes mistakes.

That system does not exist.

The strongest system is one designed to detect and correct its own mistakes before those mistakes become institutional truth.

So perhaps:

🚨 WHO INVESTIGATES THE INVESTIGATORS?

was never quite the right question.

The better question is:

🚨 WHO MAKES SURE THE INVESTIGATORS NEVER BECOME UNINVESTIGABLE?

Complainants deserve protection.

Respondents deserve fairness.

Whistleblowers deserve safety.

Investigators deserve independence.

Organisations deserve reliable findings.

And boards deserve something more meaningful than reassurance that:

“the process was followed.”

They deserve assurance that the process itself remains worthy of trust.

Because the question is not simply whether your organisation investigates people.

THE QUESTION IS WHETHER ITS INVESTIGATION SYSTEM REMAINS INVESTIGABLE.

The moment an accountability system becomes incapable of questioning itself, it crosses an important line.

It is no longer merely exercising accountability.

🚨 IT IS EXERCISING POWER WITHOUT AN EFFECTIVE MIRROR.

And that should concern every board.

References

Ask, K. and Granhag, P.A. (2005) ‘Motivational sources of confirmation bias in criminal investigations: the need for cognitive closure’, Journal of Investigative Psychology and Offender Profiling, 2(1), pp. 43–63.

Bond, C.F. and DePaulo, B.M. (2006) ‘Accuracy of deception judgments’, Personality and Social Psychology Review, 10(3), pp. 214–234.

Carter, M., Thompson, N., Crampton, P., Morrow, G., Burford, B., Gray, C. and Illing, J. (2013) ‘Workplace bullying in the UK NHS: a questionnaire and interview study on prevalence, impact and barriers to reporting’, BMJ Open, 3(6), e002628. doi:10.1136/bmjopen-2013-002628.

MacLean, C.L. (2022) ‘Cognitive bias in workplace investigation: Problems, perspectives and proposed solutions’, Applied Ergonomics, 105, 103860. doi:10.1016/j.apergo.2022.103860.

Orifici, A. (2026) Workplace Investigations: Justice and the Legal Regulation of Employer Conduct. Oxford: Oxford University Press. doi:10.1093/9780198970019.001.0001.