Chapter 4
Escaping the Tunnel Vision of Scarcity
When people face scarcity-of money, time, or mental bandwidth-they develop "tunnel vision," focusing only on immediate problems while neglecting preventive measures. This tunneling trap affects both individuals and organizations. Researcher Anita Tucker found that nurses constantly solved immediate problems (missing security tags, towel shortages) without addressing root causes. They were professional problem-solvers but worked in a system that never improved.
Tunneling becomes self-perpetuating; the constant need for heroic fixes prevents systematic solutions. Escaping requires "slack"-dedicated time for upstream thinking, like hospital safety huddles or school Freshman Success Teams. Our brains naturally prioritize immediate threats over future ones, making upstream thinking an evolutionary "add-on utility still in beta testing." Only with children and teeth do we reliably think preventively.
Climate change exploits all our psychological weaknesses, yet the successful international response to ozone depletion shows we can create urgency for prevention. The Montreal Protocol nearly failed many times but ultimately succeeded through smart framing ("ozone hole"), addressing opponents' concerns, and creating international cooperation. The phrase "the world avoided" describes the catastrophic environmental changes prevented by this upstream effort-a goal shared by all upstream interventions seeking to prevent harm, injustice, disease, or hardship.
Chapter 5
Uniting the Right People Around the Problem
In 1997, Iceland faced a serious teenage substance abuse crisis, with 42% of 15-16 year-olds reporting being drunk in the previous month. A photo from downtown Reykjavik at 3 a.m. captured streets packed with drunk teenagers-a scene considered normal at the time. Concerned leaders launched "Drug-free Iceland," uniting researchers, policymakers, schools, parents, police, churches, sports clubs, and many others around a revolutionary approach: instead of just telling teens to "say no," they aimed to change the entire culture surrounding teenagers.
The campaign transformed teenage life through multiple coordinated strategies: enforcing "outside hours" policies with official magnets sent to all parents, providing gift cards for families to spend on sports club memberships, professionalizing coaching, and creating opportunities for "natural highs" through structured activities. By 2018, the results were dramatic: in a typical class of 40 students, those who had been drunk in the previous month dropped from 17 to just 3, daily smokers fell from 9 to 2, and cannabis users decreased from 7 to 1.
Similarly, after domestic violence victim Dorothy Giunta-Cotter was murdered by her estranged husband despite following all safety protocols, Kelly Dunne faced a crisis of faith. The tragedy exposed critical gaps in a system splintered into specialized functions-police, healthcare providers, advocates, prosecutors, and parole officers-with no one responsible for preventing homicide.
Dunne discovered Jacquelyn Campbell's "Danger Assessment" tool, which accurately predicts homicide risk through a 20-question assessment. In 2005, Dunne organized the Domestic Violence High Risk Team, bringing together 13-15 people from previously disconnected roles to review cases of women scoring highest on Campbell's assessment. This unprecedented collaboration created emergency plans for at-risk women and implemented protective measures like police "drive-bys" of victims' homes. The results have been remarkable: while there were 8 domestic violence homicides in the decade before the team formed, in the 14 years since, not a single woman has been killed in their service area.
The success of both this initiative and Chicago Public Schools' Freshman On-Track program reveals a crucial pattern: surround the problem with the right people, give them early warning, and align their efforts toward preventing specific instances of harm. Both centered on data-driven discussions about real individuals-"What are we going to do about Michael next week?" or how to protect Nicole from her abuser.
Chapter 6
Changing Systems That Determine Outcomes
Anthony Iton, educated at Johns Hopkins Medical School, was shocked by the blighted neighborhoods he encountered in East Baltimore in 1985. Years later, as director of Alameda County Public Health Department, Iton analyzed life expectancy by neighborhood and uncovered shocking disparities-up to 16 years difference between ZIP codes just 22 miles apart. Similar patterns emerged in other cities, with Cleveland showing a 23-year life expectancy gap within a 4-mile walk-"like having Sweden and Afghanistan in the same city."
These gaps couldn't be explained by health care access, AIDS, homicide rates, infant mortality, or even unhealthy behaviors. Instead, Iton discovered these neighborhoods were "incubators of chronic stress" where residents constantly struggled with housing, education, crime, employment, and nutrition. This chronic stress physically altered residents' health, leading to cardiovascular disease, diabetes, and inflammation.
Systems determine probabilities-in well-designed systems, the odds favor success, while in flawed systems, success requires defying overwhelming odds. While exceptional individuals can overcome these barriers (like the inner-city kid who gets into Harvard), we celebrate these rare successes without questioning the unfair system that made their achievement so difficult.
Well-designed systems are powerful upstream interventions. Water fluoridation, named one of the CDC's top 10 public health achievements of the 20th century, saves $20 in dental costs for every $1 spent. Auto safety improvements have reduced fatalities from 5 deaths per 100 million miles driven in 1967 to about 1 death today through countless system tweaks like high friction surface treatments on dangerous curves, which reduced crashes by 80% in Kentucky.
Many social sector organizations tacitly accept the flawed systems they're trying to address. A foundation aiming to boost financial security for low-income people offered coaching rather than addressing the underlying lack of money and opportunity. Similarly, DonorsChoose has raised $875 million for classroom supplies, but risks enabling an unjust system where teachers must fundraise for basics. While such organizations provide crucial "crutches," they should simultaneously push for system change.
Chapter 7
Finding Leverage Points for Maximum Impact
Finding the right leverage point is crucial when attempting to prevent problems in complex systems. In 2008, amid Chicago's crime wave, three colleagues founded the University of Chicago Crime Lab to build evidence-based approaches to reducing violence. Challenging common assumptions about gang violence, they analyzed 200 consecutive homicides of young men and discovered most weren't strategic gang hits but impulsive reactions to trivial disputes. "My fundamental equation," said Harold Pollack, "is a couple of young guys plus impulsivity, maybe plus alcohol, plus a gun, equals a dead body."
This insight revealed potential leverage points: moderating impulsivity, reducing alcohol consumption, or restricting gun access. The Crime Lab partnered with Youth Guidance's "Becoming a Man" (BAM) program, which taught young men to distinguish between destructive "savage" anger and constructive "warrior" responses. After an agonizing nine-month wait for results from a randomized control trial, they discovered BAM participants showed a 28% reduction in arrests and a 45% reduction in violent-crime arrests compared to the control group.
Finding effective leverage points requires immersing yourself in the problem. The Crime Lab studied medical examiner reports to understand violence patterns, similar to how Permanente Medical Group leaders examined patient deaths and discovered sepsis was causing a third of them-leading to a 60% reduction in sepsis mortality after implementing solutions.
Effective leverage points often target either risk/protective factors or specific subpopulations. The Crime Lab identified just 5,000 people (0.2% of Chicago's population) who represented 17% of homicide victims. With each gunshot injury costing society approximately $1.5 million, intensive interventions costing $22,000-$23,000 per person annually become economically viable.
The healthcare field increasingly recognizes "social determinants of health"-environmental conditions affecting wellbeing from housing to air quality. The Hackensack Meridian School of Medicine is taking a revolutionary approach by placing social determinants at the core of its curriculum. First-year students are matched with families in nearby communities, helping them achieve health-related goals while learning firsthand about the complex factors affecting wellbeing. Dr. Rocchetti believes graduates will "transform the health system from within" by understanding the "power of proximity" that Bryan Stevenson advocates: getting close to vulnerable populations to identify true leverage points for change.
Chapter 8
Creating Early Warning Systems That Work
Early warning systems can dramatically improve our ability to prevent problems before they escalate. In 2010, LinkedIn faced a 30% churn rate for its recruiting product. Rather than focusing on "saving" accounts at renewal time, Roli Saxena's team discovered they could predict churn within the first 30 days of subscription by tracking product usage. Customers who engaged with the product in that critical first month were four times more likely to renew. This insight led them to create "onboarding specialists" who didn't just train clients on features but actually helped them accomplish real work. Within two years, churn was cut in half, generating tens of millions in additional annual revenue.
Even seconds of warning can be precious in emergencies. Japan's sophisticated earthquake detection network of 3,200 seismographs can detect primary waves and issue warnings in just three seconds-enough time for businesses to shut down production, doctors to halt procedures, and people to seek shelter before shaking begins. Similarly, businesses now use predictive maintenance through IoT sensors-like "smart elevators" that transmit diagnostic data to detect gradual changes signaling potential failures.
Early warning systems can sometimes cause more harm than good. South Korea's "epidemic" of thyroid cancer illustrates this danger-mass screening detected harmless "turtle" cancers (slow-growing, non-threatening tumors) that would never have caused problems, yet led to unnecessary surgeries. The country's impressive 99.7% five-year survival rate wasn't due to medical wizardry but to treating conditions that weren't threats. Similarly, alarm fatigue undermines effectiveness-a study of five ICUs found 2.5 million monitor alarms in a month, including 187 audible alarms per bed daily.
After the 2012 Sandy Hook Elementary School massacre, survivors founded Sandy Hook Promise to prevent future tragedies. Their "Know the Signs" program trains students to recognize warning signs of potential violence: social withdrawal, fascination with firearms, aggressive behavior, and explicit threats. In 2018, they launched the Say Something Anonymous Reporting System, allowing students to submit concerns without fear of stigma. When Pennsylvania schools adopted the system in 2019, it received 615 tips in the first week alone, leading to 46 suicide interventions, drug busts, and sexual assault interventions. As Nicole Hockley, who lost her 6-year-old son Dylan at Sandy Hook, explains, their goal is to interrupt the "chain of dominoes" that leads to tragedy by creating space for intervention.
Chapter 9
Measuring Success When Nothing Happens
Measuring success in upstream interventions proves challenging because, unlike downstream work that restores a previous state, upstream prevention lacks self-evident victories. We often rely on proximate measures that may not truly reflect long-term success, leading to "ghost victories" that mask failure.
The first type of ghost victory occurs when you claim credit for improvements actually caused by broader trends-"a rising tide lifts all boats." This happened in the 1990s when crime fell across America, making police chiefs with various policing philosophies all look successful. As Jens Ludwig noted, police chiefs from this era established lucrative consulting companies, while those who served during the crack cocaine era did not, despite neither group being fully responsible for the crime trends they experienced.
Katie Choe, Boston's Public Works chief engineer, discovered a ghost victory when comparing two maps: one showing sidewalk conditions and another showing repair requests. Though her department appeared successful by their metrics-equal budget allocation across zones, square footage repaired, and 311 cases closed-the maps revealed systemic inequity. Wealthy neighborhoods generated more 311 calls and received more repairs, while low-income areas with worse sidewalks were neglected. Shockingly, 45% of repairs were performed on sidewalks already in good condition, while only 15% of repairs in poor areas actually improved conditions.
When measures become the mission, people game the system. In England, hospitals kept patients in ambulances to avoid violating four-hour wait time rules. In New York, CompStat's accountability system for crime reduction led to widespread "downgrading" of crimes. The key was using Andy Grove's "paired measures" approach: balancing quantity metrics with quality checks to prevent gaming. Any upstream effort should include "pre-gaming" - anticipating how measures might be misused.
Chapter 10
Avoiding Unintended Consequences
Systems are complicated. When we intervene upstream, we create ripple effects - always. Macquarie Island's conservation saga illustrates this perfectly: eliminating cats led to rabbit overpopulation; eliminating rabbits led to invasive weed growth. In New York City, budget cuts to tree pruning saved money for parks but cost millions in injury settlements from falling branches.
The "cobra effect"-when solutions make problems worse-lurks everywhere, from colonial India's cobra bounty that increased cobra populations to open-office plans that dramatically reduce face-to-face interactions rather than enhancing them. As systems thinker Donella Meadows advised, we must view the whole system, not just our immediate problem, and recognize that changes benefiting the whole may sometimes harm parts of the system.
For experiments to succeed, we need prompt, reliable feedback-like the blue dot on Google Maps. Without it, we're driving blindfolded. Andy Hackbarth, former RAND researcher, emphasizes: "Whatever plan you have is going to be wrong. The only way you'll know it's wrong is by having feedback mechanisms and measurement systems in place."
To avoid harmful interventions, we must ask critical questions: Has a similar intervention been tried before? Can we experiment on a small scale first? Can we create closed feedback loops for rapid improvement? Is our intervention easily reversible if it causes harm?
Upstream work demands humility. "Systems can't be controlled, but they can be designed and redesigned," Meadows notes. "We can't surge forward with certainty into a world of no surprises, but we can expect surprises and learn from them and even profit from them.... We can't control systems or figure them out. But we can dance with them!"
Chapter 11
Funding Prevention: Who Pays for What Doesn't Happen?
Prevention is better than cure-yet we consistently undervalue and underfund upstream efforts. In 1890, Professor A. Arnold Clark lamented that Battle Creek spent $9 million on physicians treating illness but only $50,000 on prevention. This same argument persists 130 years later, despite overwhelming evidence of prevention's effectiveness.
Life expectancy increased dramatically from 47.3 years in 1900 to 76.8 in 2000, primarily because we stopped children from dying early. Infectious diseases that once caused 33% of deaths now cause less than 3%, thanks to upstream efforts like improved hygiene, cleaner water, sewage systems, and vaccines. Yet public health receives just 2.5% of total healthcare spending.
The fundamental problem is that reactive efforts succeed when problems happen and are fixed, while preventive efforts succeed when nothing happens. Who will pay for what does not happen? Our fee-for-service model inherently favors reaction over prevention-we'll pay $40,000 annually for insulin but balk at $1,000 to prevent diabetes.
The Nurse-Family Partnership (NFP) exemplifies this challenge. The program pairs registered nurses with first-time, low-income mothers for regular home visits from pregnancy through the child's second birthday. It costs about $10,000 per woman but yields remarkable results: reduced preterm births, fewer childhood injuries, less child abuse, decreased criminal behavior, and reduced welfare dependency. Three major randomized controlled trials show NFP delivers at least $6.50 in benefits for every dollar invested. Despite this clear success, the program faces the "wrong pocket problem"-the entity paying for prevention (often state Medicaid) doesn't receive most of the financial benefits, which flow to other agencies and budgets.
Innovative payment models like Accountable Care Organizations (ACOs) incentivize primary care doctors to prevent hospital visits. Dr. Jonathan Lilly describes how the ACO model transformed his practice-seeing fewer patients but spending more time with each, offering evening and weekend hours, and monitoring health metrics proactively. The result: healthier patients, fewer hospitalizations, and shared savings.
Chapter 12
Going Far Upstream: The Courage to Prevent Distant Threats
For those wanting to tackle larger societal problems, three key pieces of advice emerge: Be impatient for action but patient for outcomes; recognize that macro starts with micro; and favor scoreboards over pills.
The first motto, from healthcare improvement expert Maureen Bisognano, captures the upstream paradox: change requires immediate action but often takes years to show results. Sally Herndon's anti-smoking work in tobacco-friendly North Carolina exemplifies this patience. Starting in 1990, her team spent a decade achieving smoke-free policies in just 10% of school districts. But their persistence eventually led to statewide bans in schools, hospitals, prisons, government buildings, and finally restaurants and bars by 2009. Upstream victories come "an inch at a time, and then a yard, and then a mile."
Counterintuitively, solving problems at scale often begins with understanding individual cases. The most successful upstream interventions all used "by-name lists" to track individual progress. You can't help thousands until you understand how to help one person. Getting "proximate" to problems is essential-like University of Chicago Crime Lab researchers reading 200 homicide reports to truly understand violence.
The social sector often follows a flawed "Pill Model" where interventions are treated like drugs-formulated, tested in controlled trials, then distributed unchanged. While testing is valuable, this approach hinders learning and adaptation. Instead, embrace the "Scoreboard Model" where teams own problems and receive continuous data feedback to adapt and improve.
When John Koskinen was appointed Y2K czar in 1998, he faced a classic no-win situation: if nothing happened, people would call it a waste of time and money; if systems collapsed, he'd be blamed for failing to prevent disaster. When January 1, 2000 arrived with only minor glitches, critics claimed Y2K had been overblown. But many IT leaders maintain it was a catastrophe narrowly avoided through preparation-a frustrating aspect of upstream work where success means nothing happens.
Oxford philosopher Nick Bostrom uses a metaphor of humanity pulling balls from an urn to represent technological innovation. White balls are beneficial technologies (like antibiotics), gray balls are mixed blessings (like automobiles), but Bostrom warns there might be a "black ball"-a technology that destroys the civilization that invents it. What society needs are "enlightened Chicken Littles" like Bostrom, who raise awareness about existential risks without exploiting fear.
Upstream thinking isn't just for organizations but for individuals confronting recurring problems. Examples include a woman who eliminated parking stress by deliberately choosing remote "VIP spots," tennis coach Jake Stap who invented the tennis ball hopper to solve his back pain from bending, and a couple who created "hot tub conversations" to safely discuss difficult topics and save their marriage.
The best time to plant a tree was 20 years ago. The second-best time is now.
