LEADERSHIPOngoing practice

Naive Interventionism and Iatrogenics

First, do no harm -- the cost of intervention often exceeds the cost of inaction

Problem it solves

ineffective leadership

Best for

["managers tempted to micromanage","policymakers evaluating interventions","healthcare consumers navigating treatment options","anyone with a bias toward action when inaction may be wiser"]

Not ideal for

["genuine emergencies requiring immediate decisive action","situations of clear, severe pathology where intervention is clearly warranted","contexts where the cost of inaction is demonstrably catastrophic"]

Overview

Why this framework exists

Iatrogenics -- harm caused by the healer -- is Taleb's term for the damage inflicted by well-intentioned interventions. The concept originates in medicine (where until penicillin, doctors often decreased patient life expectancy) but generalizes to every domain: economics, education, parenting, governance, foreign policy. Whenever someone intervenes in a complex system, the intervention has both intended and unintended consequences. When the unintended harm exceeds the intended benefit, we have iatrogenics.

Naive interventionism is the compulsion to 'do something' in response to any perceived problem, without adequately weighing the cost of the intervention itself. It is driven by a asymmetry in incentives: it is easier to sell 'look what I did for you' than 'look what I avoided for you.' The doctor who operates gets paid and praised; the doctor who wisely refrains from unnecessary surgery is invisible.

Taleb argues that complex systems have self-healing properties that interventionists systematically ignore. The body heals itself after most injuries; economies recover from most recessions; children learn from most mistakes. The burden of proof should be on the intervenor to demonstrate that their action will do more good than harm, not on the system to prove it can heal itself. The Fabian strategy -- named after the Roman general who defeated Hannibal by refusing to engage -- shows that strategic inaction can be the most powerful intervention of all.

Core principles

7 total
  1. Intervention in complex systems produces unintended consequences that often exceed intended benefits
  2. The burden of proof is on the intervenor, not on the natural system
  3. It is easier to sell action than inaction, creating a systematic bias toward intervention
  4. Complex systems have self-healing properties that interventionists ignore
  5. Noise (random fluctuations) is constantly confused with signal (meaningful information), triggering unnecessary interventions
  6. The more frequently you check data, the more noise you see, and the more you intervene from noise
  7. Procrastination (the Fabian kind) is a natural heuristic for filtering out noise

Steps

5 steps
  1. Apply the Break-Even Test
    Before any intervention, estimate both the benefits and the costs (including second-order effects, risks of the intervention itself, and opportunity costs). If the expected net benefit is not clearly positive, do not intervene. Remember that the intervention has risks too -- the operation itself can kill.
  2. Distinguish Noise from Signal
    Reduce the frequency with which you check data and metrics. The more frequently you observe a system, the higher the ratio of noise to signal, and the more likely you are to intervene on randomness. Check quarterly instead of daily. Review annually instead of monthly. Use the Fabian principle: time is the best filter.
  3. Impose a Waiting Period Before Action
    When you feel the urge to intervene, impose a mandatory waiting period. Many perceived problems resolve themselves. Many 'urgent' situations are not actually urgent. Strategic procrastination (waiting to see if the problem persists before acting) filters out noise and allows natural self-correction to operate.
  4. Reserve Intervention for Severe, Clear-Cut Cases
    Intervene only when the condition is severe and the evidence for intervention is strong. In medicine: operate on life-threatening conditions, not mild ailments. In management: address existential threats, not performance fluctuations. In parenting: prevent genuine danger, not normal childhood struggle. The more severe the condition, the more likely the intervention's benefits exceed its costs.
  5. Make Non-Intervention Visible and Valued
    Change incentive structures to reward avoidance of unnecessary intervention. Celebrate the doctor who did not operate, the manager who did not reorganize, the policymaker who did not regulate. The hero who prevents a disaster by inaction deserves as much recognition as the one who responds to one.

Examples

1 cases
The Tonsillectomy Study

In the 1930s, 389 children were presented to New York City doctors; 174 were recommended for tonsillectomies. The remaining 215 were shown to different doctors, and 99 more were recommended for surgery. The remaining 116 were shown to yet another group, and 52 more were recommended. Each unnecessary surgery carried 2-4% morbidity risk and approximately 1 in 15,000 chance of death.

OutcomeThe study revealed systematic overintervention driven by the bias toward action. Doctors consistently recommended surgery regardless of whether it was needed, because the incentive structure rewarded doing something over doing nothing. The probabilistic harm to children from unnecessary operations was invisible while the visible 'help' of surgery was rewarded.

Common mistakes

3 traps
Equating Inaction with Passivity
Not intervening is itself an active decision that requires courage, judgment, and willingness to absorb criticism. Fabian's refusal to engage Hannibal was ridiculed by Romans who wanted decisive battle. But strategic non-intervention is not doing nothing -- it is choosing the more difficult path of patience when the easy path is action.
Applying Non-Intervention to Acute Emergencies
Iatrogenics is most prevalent in mild, ambiguous situations where the system can self-heal. In genuine emergencies (severe trauma, existential business threats, imminent danger), intervention is warranted because the cost of inaction is catastrophic and certain. The framework applies to the spectrum, not the extremes.
Checking Data Too Frequently
Looking at daily stock prices, hourly website metrics, or moment-by-moment performance data floods you with noise, triggering interventions that are harmful on average. Taleb demonstrates that the ratio of noise to signal increases dramatically with observation frequency. Check less often, and you will intervene less, and your system will perform better.

Origin story

How this framework came to be

Taleb draws on the history of medicine, where harmful interventions were standard practice for millennia. George Washington's doctors hastened his death through bloodletting. Ignaz Semmelweis discovered that doctors were killing patients with unwashed hands but was committed to an asylum for challenging the establishment. Hospital fever killed more women in childbirth than giving birth on the street. The insight generalized when Taleb observed the same pattern in economics (Federal Reserve interventions amplifying crises), education (over-testing reducing learning), and parenting (helicopter parenting creating fragile children).

Source

Traced to primary
Source · BOOK
Antifragile: Things That Gain from Disorder
Nassim Nicholas Taleb · 2012
Open source →

Related frameworks

Browse all Leadership →