Interactive exercise companion to The Formula for Better Health teaching resources by Dr. Tom Frieden. This tool is designed to be used alongside the original teaching materials — visit formulateaching.theformulaforbetterhealth.net for the full instructor resources.
The Formula for Better Health
Chapters 8 & 1 · Smoke-Free New York · Interactive Exercise
SEE
BELIEVE
CREATE
Step 1 of 4 — The political landscape
SEEBELIEVECREATE
New York City, 2001–2003
The health evidence is unambiguous. Passing the law is not.
"Thunder is good, thunder is impressive, but it's lightning that does the work."
— Mark Twain, quoted in the case
What this exercise is for
New York City's Smoke-Free Air Act — banning smoking in all workplaces, restaurants, and bars — required more than strong evidence and political backing. It required rigorous political strategy. This exercise walks you through four critical decision points in that campaign, then asks you to confront what happened after the law passed.
Part 1 covers the political fight (CREATE). Part 2 covers what the data revealed once smoking began to decline — and then stall (SEE). Use the phase navigation above to move between parts at any time.
Graduate framing
The instructor guide sets a core analytical challenge: students will invoke "political will" as an explanation for the law's success. Your task throughout Part 1 is to replace that explanation with specific mechanisms. What precisely did the campaign do differently? What structural conditions made those mechanisms work?
Step 2 of 4 — Winners and losers
Framework: Winners & Losers
Every public health policy creates concentrated losers and diffuse winners. Map them.
Graduate framing
The asymmetry between concentrated losers and diffuse winners is the central political problem in public health. Losers have motive, money, and organisation. Winners don't yet know they've won, and their benefits are years or decades away. Notice as you map these stakeholders: the same population can appear on both sides depending on timeframe.
Assign each stakeholder to the correct group. Click a tag to place it; click a placed tag to remove it and reassign.
Please assign all five stakeholders before checking.
The soda tax faced the same structural problem — but more severely. The beverage industry was bigger and better organised, and the winners were no more structured than in the smoke-free case. Same mayor, same evidence, same city — different political conditions, different outcome. Evidence does not determine political outcomes. Political conditions do.
Two decisions defined the campaign's political architecture.
Graduate framing
Both decisions involve the same underlying insight: political costs are not fixed — they depend on when in the process you act. Pre-election, the cost of commitment is low; post-election, reversal is costly. Early in an administration, political capital peaks; later it has depreciated. Timing is structural leverage, not tactical convenience.
Decision A — Joe Cherner's pre-election strategy
Before the 2001 election, Cherner sent detailed written questionnaires to every mayoral and City Council candidate requesting their position on smoke-free workplaces. Bloomberg and most Council members signed written commitments. Cherner published every response publicly.
Why did this matter more than lobbying elected officials after the law was proposed?
Compare this to the soda tax: by the time sustained advocacy began at the state level, legislators were already facing intense, organised industry opposition with no equivalent prior commitment on record. Cherner's questionnaire required someone willing to do organising work before any bill existed — which is why it is rarely replicated despite being replicable.
Decision B — The smoking room compromise
The City Council speaker refused to pass the bill without allowing separate smoking rooms. Actual smoking rooms would have undermined the law and made a subsequent, stricter law nearly impossible to pass. The health department's solution: permit smoking rooms — but require tuberculosis isolation-level engineering specifications so demanding that building one was prohibitively expensive. Two bars investigated; neither built a room.
What principle does this illustrate?
This requires technical knowledge in service of political strategy — knowing that tuberculosis isolation-level specifications would make rooms impractical required public health expertise, not political instinct alone. Political competence and technical competence are not separate; the most effective political solutions in public health often require both simultaneously.
Please answer both questions before continuing.
Step 4 of 4 — Advocates, partnerships, and timing
Framework: Advocates & Partnerships · Do the Hard Things First
The law passed. Why — and what principle governs the timing?
Graduate framing
The chapter's synthesis: political skill is a form of competence — not character, not charisma, not will. It is teachable and learnable. As you work through this step, identify the specific mechanism behind each element of the campaign's success. Resist any answer that names a variable without describing a mechanism.
The most persuasive witness
At the City Council hearing, speakers included Mayor Bloomberg, Nobel Laureate Harold Varmus, and then-Health Commissioner Frieden. City Council leaders later reported that the most persuasive testimony came from Martinah Payne-Yehuda, a pregnant waitress who worked in a smoking bar. Why?
This does not mean institutional testimony is unimportant — Bloomberg's commitment, Varmus's scientific authority, and Frieden's public health expertise each served a distinct function. The lesson is about coalition design: the triad of strong government, proactive civil society, and rigorous monitoring needs a fourth element no institution can supply — the face and voice of the person the law is designed to protect.
"Do the hard things first"
Bloomberg implemented the Smoke-Free Air Act, tobacco tax increases, and syringe exchange programmes in his first term. Why does the timing of a new administration matter so much to which interventions are politically possible?
This principle has a corollary: interventions that cannot be implemented in year one may effectively be off the table until the next election. A commissioner who waits for the "right moment" later in a term may find it never arrives. The incoming health commissioner's 90-day priority question is therefore not "what is most important" but "what is most politically executable now, with the capital available now."
Please answer both questions before continuing.
CREATE→SEE
The law passed. Now what did the data show?
The Smoke-Free Air Act came into force in 2003. The political strategy had worked. Smoking fell sharply. But then something the citywide numbers almost concealed happened: the decline stalled. What did the surveillance system reveal — and was it designed to reveal it?
140K
fewer smokers, Year 1
21.6→19.2%
adult smoking rate
then…
the decline stalled
Part 2 shifts from political strategy to programme surveillance. The central question: when a programme stops working, how does anyone know — and what must be in place to find out and respond?
Part 2 · Step 1 of 3 — Two kinds of surveillance
SEEBELIEVECREATE
Surveillance for programme management
The law worked. Then smoking stalled. The question is: how did anyone know?
Graduate framing
Chapter 1's surveillance definition ends with "dissemination to those who need to know." That phrase contains a management theory. Data flowing only upward to policymakers once a year fulfils documentation. Data flowing to the programme manager who can redirect outreach workers weekly fulfils management. Who receives the data, at what granularity, with what frequency, and what decision does it trigger — these are as important as the measurement itself.
Chapter 1 distinguishes two kinds of surveillance. Assign each card to the type of surveillance it describes. Click a placed card to return it to the pool and reassign.
?
Where is the disease? Who is at risk? What is its transmission pattern? (The question that traced MDR-TB chains and projected the Ebola epidemic arc.)
?
Is the intervention working? Where is it succeeding? Where is it falling short? (The question the Community Health Survey was designed to answer.)
?
Has the programme reached the neighbourhoods where burden is highest? (Aggregate citywide data cannot answer this. Neighbourhood-level data can.)
?
Is this cluster of cases linked? Is the outbreak spreading or contained? (The question that initiates an emergency response.)
Disease detection
Assign cards here
Programme management
Assign cards here
Assign this card to:
A surveillance system that detects a problem but reports it to someone who cannot act on it is not a management tool. The same data, flowing to different people at different granularity, produces different organisational outcomes. Building that data flow into programme design is as important as building the measurement system itself.
Please assign all four cards before checking.
Part 2 · Step 2 of 3 — The feedback loop
Detect → Diagnose → Intervene → Remeasure
Citywide smoking rates plateaued. The Community Health Survey revealed what the aggregate data hid.
Graduate framing
Aggregate success can mask persistent inequity. A programme measuring only average improvement implicitly defines success as average improvement — allowing the most disadvantaged populations to be overlooked systematically. The NYC stall illustrates this: citywide rates appeared to plateau, but the plateau was the product of continued decline in some neighbourhoods and a complete stall in others. Without neighbourhood-level disaggregation, the equity dimension is invisible.
The stall was detected and corrected through a four-step feedback loop. Click each step in the order it occurred (1 → 4). Click a selected step again to deselect it and correct your sequence.
?
Data matched against programme reach — standard campaigns not breaking through in lower-income neighbourhoods
Click to deselect and reassign
?
Continued annual monitoring confirmed whether the redirect was working
Click to deselect and reassign
?
Community Health Survey detected the stall and located it by neighbourhood
Click to deselect and reassign
?
Outreach redirected — hard-hitting campaigns targeted at specific neighbourhoods; tobacco taxes continued to rise
Click to deselect and reassign
The tax mechanism is worth examining specifically: cigarette taxes reduce smoking most powerfully among lower-income populations, who are more price-sensitive. As New York's prices climbed among the highest in the country, the additional reductions were concentrated in exactly the communities the standard campaigns had not reached. The feedback loop did not just detect a problem — it revealed which tools would reach the populations being left behind.
Please order all four steps before checking.
Part 2 · Step 3 of 3 — Course correction and debrief
Three conditions for course correction
Detecting a stall is necessary. It is not sufficient. What else must be in place?
Graduate framing
Most students identify the information system as the condition most often absent. Push beyond this: data that reveals failure is institutionally threatening. Programmes have inertia. The people who designed them have professional investment in their success. Willingness to acknowledge failure — and the leadership courage to change course publicly — may be rarer than the surveillance capacity to detect it.
Three conditions must all be present for a programme to adjust based on performance. Click a card to select it, then choose which condition it describes. Click a placed card to return it and reassign.
The Community Health Survey — neighbourhood-level smoking rates, tracked annually, reported to programme managers who could act on the results.
NYC's readiness to conclude that standard campaigns were insufficient in certain neighbourhoods, resume hard-hitting advertising, and change strategy publicly.
The capacity to redirect outreach workers, produce new campaign materials, and sustain the change long enough to measure its effect.
Assign this card to:
1. Information system
Select a card above, then assign here
2. Willingness to acknowledge failure and change course
Select a card above, then assign here
3. Financial and human resources to implement the correction
Select a card above, then assign here
The second condition — willingness to acknowledge failure — is worth pausing on. Data that reveals a programme is not working is institutionally threatening. Unfavourable findings are often suppressed, reframed, or not acted upon. NYC's willingness to name the stall publicly and change course required leadership willing to accept the political cost of admitting that the initial approach had not been enough in certain communities.
Please assign all three cards before checking.
Debrief
Bringing it together: political competence and programme surveillance
The smoke-free law succeeded because it applied a rigorous political strategy — not because Bloomberg had unusual "political will." Which specific mechanism in the campaign is hardest to replicate in your own context, and why?
Consider: Cherner's pre-election survey required someone willing to organise before any bill existed. Payne-Yehuda's testimony required identifying and supporting an affected individual willing to speak publicly. The engineering specifications required technical knowledge deployed in service of political strategy. Which of these is least available in the systems you know?
The stall was visible only because the Community Health Survey was designed with neighbourhood-level granularity. What level of disaggregation would your programme require to detect a similar failure — and who currently receives that data?
The right level of disaggregation is determined by the level at which interventions can be redirected — not by what is easiest to measure. If a programme can redirect outreach at the district level but measures only nationally, the data cannot trigger the decision it needs to trigger.
Which of the three course correction conditions — information system, willingness to change, or resources — is most often absent in programmes you know?
Most students identify the information system. Push further: are there cases where the data existed but was not acted upon? What does that suggest about where the binding constraint actually lies?
The broader lesson
Political competence and programme surveillance are both forms of the same discipline: designing systems that close the loop between action and evidence.
CREATE
Political skill is a form of competence. Identifying winners and losers, organising winners, reaching deciders before they are locked in, building coalitions that make diffuse benefits visible, and finding solutions that meet political constraints without sacrificing health outcomes — each of these is teachable and learnable.
SEE
A programme that stops working will continue to consume resources invisibly unless it includes a feedback mechanism designed to detect and respond to failure. Data is a management tool only when it is designed to change operational decisions — at the right granularity, with the right frequency, reaching the person who can act on it.
New York City's tobacco campaign succeeded because it did both: it won the political fight with rigorous strategy, and it used the Community Health Survey to detect when the decline had stalled — and to guide the response that got it moving again.
Interactive exercise design, scenario architecture, and tool format by Dr. Louisa Sun, National University Health System. Licensed under CC BY-NC-SA 4.0.