Writing for Policy

Science is not finished when it is published; it is finished when a decision-maker can use it. Most academic writing never gets that far, because it is built for a different reader with a different job. This page is the companion to scientific writing: where that page teaches how to state a hypothesis, frame Specific Aims, and structure a manuscript, this one teaches how to turn a result into something a health policymaker will actually read and act on.

The spine here is Christopher Whitty’s account of what makes an academic paper useful for health policy, written as he stepped down as Chief Scientific Adviser at the UK Department for International Development.

Whitty CJM. What makes an academic paper useful for health policy? BMC Medicine. 2015;13:301. doi:10.1186/s12916-015-0544-8

Why policy writing is different#

A journal reader and a policymaker want different things from the same result. The journal reader wants to know whether the work is correct and where it sits in the literature. The policymaker wants to know what decision the work bears on and whether it changes the answer.

Four differences follow from that.

• Audience. The reader is senior, numerate, and not a specialist in your field. They are comfortable with effect sizes, ranges, and opportunity cost, but not with the jargon of your subfield. Many are trained as economists or advised by them, so the question they always ask is what a choice costs and what it displaces.
• Purpose. The paper exists to inform a specific decision, not to advance a literature. A result with no decision attached has no policy use, however good the science.
• Timeline. Policy moves in days to weeks, not the months of a review cycle. Evidence has to be assembled before the decision window opens, because there is no time to commission it once the window is open.
• The cost of length and hedging. A reader who will stop after the first hundred words cannot afford your throat-clearing. Every qualification that protects you in peer review costs you clarity with a policymaker, so hedging has to be deliberate and load-bearing, not reflexive.

Whitty’s criteria for a useful policy input#

Whitty lists properties that separate a useful policy paper from an unusable one. Each turns into a checklist item you can hold a draft against.

• State the policy problem first. A policy problem is not a scientific problem. It usually contains several scientific sub-questions across different disciplines, and naming it is the writer’s job, not the reader’s.
• Put methods and limitations on the page. Rigor for a policy audience is at least equal to rigor for a journal, not less. A limitation that is buried will resurface in the meeting, so state it plainly and move on.
• Be as simple as possible, but no simpler. Simplicity is respect for the reader’s time, not a concession on accuracy.
• Synthesize before you recommend. Single studies should rarely drive policy. The highest-value contribution is an honest synthesis of the relevant evidence, including the qualitative, observational, and economic work that a trials-only view would discard.
• Make the economics visible. Surface cost, cost-effectiveness, and opportunity cost even when the paper is not primarily economic. If no costed analysis exists, say so and name it as a gap.
• Let policymakers vary the assumptions. A transparent model whose inputs can be adjusted is more useful than a black box that emits one number.
• Do not append a “policy implications” section by reflex. Scientists tend to overestimate the reach of their own findings and to bend the analysis toward a tidy story. If implications belong, keep them short, conditional, and explicit about what is out of scope.

The single mistake Whitty warns against most sharply is the distinguished scientist offering an evidence-light opinion. It is no more useful than any other advocacy, and it adds to the noise a policymaker is already filtering.

The shape of a policy brief#

A policy brief is about two pages, roughly a thousand to fifteen hundred words, and it opens with the answer. The convention is bottom-line-up-front (BLUF): a short block at the top that a reader can stop after and still be correctly informed.

Bottom line

Policy problem. One sentence naming the decision in scope and who makes it.

State of evidence. One or two sentences synthesizing what the evidence supports.

Principal uncertainty. One sentence naming what is most likely to change the answer.

After the BLUF block the body expands, with headings that state findings rather than label sections. “Cost-effectiveness of weekly testing” beats “Results,” because a policymaker skims the headings before reading a word.

The policy problem — the decision, the actor, the constraint, the timeframe, and the scientific sub-questions inside it.

What the evidence shows — synthesis across the relevant studies, with a table of effect, cost, and uncertainty where the analysis allows.

Cost and opportunity cost — what the option costs and what it displaces.

What is uncertain — the specific thing that would change the answer, not a generic call for more research.

What this brief does not address — political feasibility, distribution, implementation cost, or whatever is out of scope, named so the brief cannot drift into advocacy.

Presenting uncertainty so it informs rather than paralyzes#

Uncertainty stated as a shrug (“more research is needed”) is useless, because every decision is made under uncertainty and the reader knows it. Uncertainty is useful when it is specific and conditional. Name what you do not know, say how much it would move the answer, and, if new evidence would resolve it, say what kind, on what timeline, and at what cost. Conditional phrasing carries this well: if the priority is speed, the evidence supports option A; if the priority is cost, the evidence is inconclusive.

The brief is one of several formats#

The two-page brief is the default, but it is not the only policy output, and matching the format to the decision matters.

• Policy brief — about two pages, for a decision that is imminent and narrow.
• Policy paper — four to eight pages, when the decision needs the reasoning laid out, not just the answer.
• Synthesis review — a policy-facing evidence review that pulls together the relevant studies across disciplines; this is the highest-value contribution Whitty names.
• White paper — a longer document that sets out a position and its supporting evidence for a slower, more strategic decision.
• Problem statement — a single page that frames the decision and names what evidence would resolve it, useful when the science does not yet exist.

Whichever format you choose, the openings share the same discipline: state the decision first, synthesize rather than advocate, and be explicit about what is uncertain and what is out of scope.

Worked example: from manuscript to brief#

The scientific writing page builds a set of Specific Aims around pre-symptomatic transmission, ending in a comparison of symptom-based versus test-based isolation. Here is how the same result reads first as a manuscript paragraph and then as a policy-brief paragraph.

A manuscript Results paragraph reports what happened, led by the numbers, without interpretation.

We estimate that 54% (95% CrI 41–66%) of transmission occurred before symptom onset. Embedding this estimate in a renewal-equation model, symptom-based isolation reduced the effective reproduction number $R_t$ from 2.1 to 1.4, leaving $R_t > 1$ across the plausible range of the pre-symptomatic fraction. Test-based isolation with a two-day turnaround reduced $R_t$ to 0.8.

The same result reframed for a brief leads with the decision, converts $R_t$ into plain consequence, surfaces cost, and states the uncertainty that would change the answer.

Policy problem. Whether to fund rapid testing for isolation release, or continue to isolate on symptoms alone, during the next respiratory-pathogen wave.

More than half of transmission occurs before symptoms appear, so isolating people only once they feel ill misses the window when they are most infectious. In our model, symptom-based isolation slows spread but does not stop it: each case still infects more than one other person on average. Testing people and releasing them on a negative result cuts transmission below the level needed for the outbreak to shrink. The result turns on how fast tests return: a two-day turnaround works, and the benefit erodes as turnaround lengthens. A costed comparison of testing capacity against the cases averted is the missing piece and should be commissioned before the decision.

The manuscript paragraph is correct and complete for a journal. The brief paragraph is what lets a decision-maker act, because it names the decision, translates the statistic into a consequence, makes the cost question explicit, and is honest about the one input that governs the answer.

A checklist before you send#

• Is the policy problem in the first paragraph, with the actor and the timeframe named?
• Could a numerate non-specialist follow it without your field’s jargon?
• Are methods and limitations on the page rather than buried?
• Is this synthesis across the evidence, or a single study standing in for it?
• Is the economic dimension visible, or the gap named?
• If a model is used, can the reader vary its assumptions?
• Has the implications section been cut or kept short, conditional, and modest?
• Could it have been written, sent, and read inside a two-week window?

Why it matters#

The evidence that reaches a decision is not always the best evidence; often it is the evidence that arrived in a form the decision could use. Learning to write for policy is how a researcher makes sure their work is in that second category. It is the same discipline as the rest of scientific writing turned toward a different reader: state the problem sharply, show your methods, be honest about what you do not know, and respect the time of the person who has to decide.

Related#

• Scientific Writing — hypotheses, Specific Aims, and manuscripts
• Scientific Pathways — the research career and grant writing
• Cover Letters and Reviewer Responses
• Scientific & Policy Writing — the full hub
• People, Plagues, and Policy — the course that assigns this page
• Risk communication and community engagement
• Reproduction number $R_t$
• Cost-effectiveness analysis