On-Screen Questions on Harm Reduction Vending Machines: What to Ask and What Not To

On-screen questions on a harm reduction vending machine turn a supply point into an evaluation tool. The principle is ask, record, never block: the machine may ask a short question before it releases supplies, but every question can be skipped, and no answer stops the release. This page covers what to ask and why a funder wants it, whether to ask on the machine or on the user’s phone, how many questions people will tolerate, what never to collect, and how the answers reach the funder’s report.

Quick answer

• Ask one to three questions per visit, as buttons, each with a prefer-not-to-say option.
• The useful questions: first supply or refill; was the last naloxone used; age band; postcode or area; how the person heard about the machine.
• Never block. A question records information; it is not an eligibility gate.
• Never collect names, full dates of birth, phone numbers on the machine, images or free text.
• Touchscreen for quick questions, phone for anything longer, with a QR code to an optional follow-up survey.
• Answers sit on the dispense record and are reported in aggregate, with a set retention period.

Why ask anything at all?

Because the funder pays for evidence, and a dispense count alone is thin evidence. A machine that gave out 600 naloxone kits last quarter says nothing about whether those kits reached people at risk, or were used. A machine that can also show how many were refills after a kit was used on someone tells a far stronger story.

Government buyers already ask for this. Victoria’s tender for naloxone dispensing machines (it closed in November 2024) listed data requirements that, as summarised in our harm reduction vending post, included user-reported data on initial supply versus refill and previous naloxone use outcome, “randomly assigned demographic questions”, and QR-code links to optional web-based surveys. The tender documents sit behind the state’s tender portal, so treat that wording as our summary rather than the official text.

The risk runs the other way too. A 2023 scoping review of harm reduction vending machines found anonymity was among the reasons people used them, and that some feared using machines in public because of police or a lack of anonymity (Russell et al., 2023). Every question spends some of that trust.

Ask, record, never block

Three rules keep questions on the right side of that line.

1. Every question is optional. Each screen has a skip button or a prefer-not-to-say answer, and skipping is recorded as a valid answer, not an error.
2. No answer stops the release. The NSW needle and syringe program guideline says participation in “brief interventions, research and evaluation surveys” must be “based on the client’s informed and voluntary consent”, and that services should avoid “unwanted educational or referral interventions which may discourage future access”.
3. Age is recorded, not enforced. Victoria’s tender, as quoted in our post, required the machine to “confirm the recipient is 12 years of age or over (but not prevent access)”. The Victorian Department of Health describes naloxone as “suitable for people of all ages”. The question exists so the program can report how many young people use the machine, and so staff can follow up the pattern, not so a 15-year-old is sent away from a medicine that reverses an overdose.

If a program genuinely needs an eligibility check, for example a quota tied to a registered code, call it what it is and put it in the release design, not the survey. Our controlled dispensing guide covers codes, cards and quotas. Keep the two separate on the screen, so users can see which steps are required and which are not.

Touchscreen or phone?

Where the question is asked	Good for	Weak at	Hardware
Machine touchscreen, before release	One or two quick questions from every user; no phone needed	Privacy (people nearby can see); time at the machine	Optional 7–32 inch Android touchscreen on our compact wall-mount machines
User’s phone, at registration	Longer questions, eligibility, consent text, languages	Only reaches people with a phone and data; one-off	None on the machine; the machine can run screen-free with a QR code
User’s phone, after release via QR	Follow-up questions, experience, outcomes	Low response rate	A QR code on the panel or the delivery screen

On our public health builds, registration happens on a web page on the user’s phone. On HIVConnect the user answers a few questions and provides a mobile number, receives an SMS code, collects at the machine, and is offered an optional follow-up survey. On MyTest the user answers a few simple questions when registering. In both, no personal data lives on the machine. A free sanitary products client of ours gates free dispensing behind a questionnaire on the phone, to verify eligibility and capture usage information.

For harm reduction supplies, a common split is one question on the touchscreen at the point of release, and a QR code for anything longer.

Which questions to ask

Each row below is an example, not a template. Use the questions your funder will actually report on.

Question	Why a funder wants it	Answer buttons	Privacy risk
Is this your first supply from this machine, or a refill?	Reach versus retention; new users are the headline outcome	First time / Refill / Prefer not to say	Low
(If refill) Since your last kit, did you use naloxone on someone?	The closest a machine gets to an outcome measure	Yes / No / Kit lost or given away / Prefer not to say	Medium: describes drug-related events; keep in aggregate
(If used) Did the person recover?	Overdose reversals reported by users	Yes / No / Not sure / Prefer not to say	Medium to high: sensitive; consider asking only on the phone survey
Who is this for?	Shows secondary distribution to friends and family	Me / Someone I know / I work or volunteer in a service / Prefer not to say	Low
Which age group are you in?	Reach to young people and older users; records 12+ without blocking	Under 12 / 12–17 / 18–24 / 25–34 / 35–49 / 50+ / Prefer not to say	Low with bands; higher with exact age
Gender	Equity reporting	Woman / Man / Another gender / Prefer not to say	Low
What is your postcode or ZIP code?	Shows where users travel from; supports siting decisions	Number pad, plus Prefer not to say	Medium: combined with age, gender and time, can narrow to one person in a small area
How did you hear about this machine?	Tells the program which promotion works	Service staff / Friend / Sign or poster / Online / Prefer not to say	Low

Two notes on wording. Use plain language at a low reading level, and avoid stigmatising terms: “use naloxone on someone”, not “witness an overdose event”. And put the most important question first, because many people will answer one and skip the rest.

How many questions, and how long?

People at a harm reduction machine are often in a hurry, sometimes late at night, sometimes on behalf of someone else. A 12-year study of Paris syringe machines found it took 14 seconds for a person to collect a syringe (Duplessy and Reynaud, 2014). Questions that double that time will be skipped or gamed.

Melbourne is the warning. At four syringe dispensing machines from 2017 to 2020, users had to enter age category, gender and postcode on a 0–9 numeric keypad before every order. Researchers estimated that “approximately 50% of all SDM orders inputted invalid demographic data”, and 44,195 orders (24% of 180,989) showed obvious repeated patterns, the same number entered in every field. Mandatory questions, asked every time, produced data that could not be trusted.

What works better:

• One to three questions per visit, as a working budget.
• Buttons with words or icons, not number codes the user has to look up.
• Rotation. Put five or six questions in a pool and show each user one or two, chosen at random. Every question still gathers enough answers across hundreds of visits, and no one faces the full set. Victoria’s tender asked for “randomly assigned demographic questions” for the same reason, as summarised in our post.
• Ask the slow-changing questions once. Age band and gender do not change between visits. Where a code identifies a returning user, ask them at registration on the phone, not at every release.
• Check quality. Flag answer patterns that look like button-mashing, and report the skip rate alongside the answers.

Skip, prefer not to say, and honest data

A visible skip option improves data. When people can skip, the answers they do give are more likely to be true. Record “prefer not to say” and “skipped” as their own values, and report them. A funder that sees a 30% skip rate on gender understands the data; a funder that sees invented answers does not.

Accessibility and languages

• Pictures as well as words. Program staff in a rural Kentucky study pointed out that “somebody may not be able to read and write so a touchscreen with images is gonna be beneficial” (Marschke et al., 2025).
• Languages. Our compact wall-mount machines can be specified with a multi-language interface. New York City’s machines post instructions in multiple languages, and the MyTest kits carry instructions in English with translations in nine languages. Choose the languages from the community the machine serves.
• Large buttons and high contrast, readable at night and under poor lighting.
• Reach. NSW asks for dispensing chutes to be sited at a height that deters access by children. That can conflict with reach for a wheelchair user, so test the screen height on site before mounting.
• A way out. A “skip all” button on the first screen respects people who just want supplies.

What not to collect

Do not collect	Why
Names	Not needed to count or evaluate; deters use
Full date of birth	An age band answers the funder’s question; a date of birth helps identify a person
Phone numbers or email typed on the machine	Visible to others and stored on a public device; collect on the user’s own phone if a quota needs it
Photos, video or fingerprints	People who use drugs have described fear that police could access fingerprint records; cameras deter use
Free text	Invites personal detail, is hard to report and needs moderation
Exact addresses	A postcode or area is enough for siting decisions

NSW’s guideline is blunt: “The collection of personal information is not required for the provision of injecting equipment.”

Data minimisation, retention and who sees the data

Privacy law already points the same way.

• Australia. Australian Privacy Principle 2 says individuals must have the option of not identifying themselves or using a pseudonym. APP 3 limits collection to what is “reasonably necessary” for an entity’s functions, and APP 11.2 requires information no longer needed to be destroyed or de-identified.
• UK and EU. The UK GDPR data minimisation principle requires personal data to be “adequate, relevant and limited to what is necessary”, and the ICO adds: “You must not collect personal data on the off-chance that it might be useful in the future.”

In practice:

1. Keep identity off the machine. On our builds the machine sees a code or token, never a name.
2. Keep answers separate from contact details. If a registration service holds a phone number for a quota, do not join it to the survey answers unless the program’s rule requires it.
3. Set a retention period for raw answers, and keep only aggregate tables after it.
4. Limit access. Name who can view and export raw records; give everyone else the aggregate report.
5. Suppress small numbers in published reports, so a row with one or two users at a small site cannot be traced back.

This is a summary for planning, not legal advice. Rules change; confirm with your privacy regulator and funder. Facts stated as of October 2026.

The optional follow-up survey

Some questions are too sensitive or too long for a machine: experiences of overdose, use of services, whether the person would like a call back. Put them on an optional web survey reached by a QR code on the panel or on the screen after release. HIVConnect does this with an optional follow-up survey. Keep it short, say who runs it and how the data is used, and do not tie it to the dispense. People in the Kentucky study unanimously favoured a callback feature, where users choose to request contact from a service, over the service chasing them; a QR link to a callback request does that without putting a phone number on the machine.

How answers reach the funder’s report

Report line	Where it comes from
Units by product, site, day and hour	The order record each release writes: items, quantities, machine, location, time, order state, receipt ID
First supply versus refill; naloxone use reported	On-screen answers stored against the order record
Age bands, gender, areas served	On-screen or registration answers, in aggregate
Skip rates and data quality	Count of skipped and prefer-not-to-say answers
Follow-up outcomes	The separate QR survey, reported on its own
Stock-outs, downtime and faults	Machine telemetry on the cloud dashboard

The cloud dashboard exports records by date range for the program’s own analysis. Agree the report fields with the funder before launch and design the questions backwards from them. The evidence base for these machines is still thin (see do harm reduction vending machines work?), so good, honest program data is worth having.

What to specify if you are buying one

1. The questions, answer options and order, and which ones rotate.
2. Touchscreen size, or screen-free with questions on the phone.
3. Languages and any icon set.
4. Retention period, export rights and who sees raw records.
5. The report the funder expects, field by field.
6. The follow-up survey link and who runs it.

Start with the compact wall-mount machine, published at about US$800–1,500 per unit, with the touchscreen as an option. A code-and-quota system with a registration web page, like our monthly-quota project, is published at US$11,000 to build and about US$200 a month. We build the question flow to the requirement you state; we do not certify compliance with any privacy law. Support is by email, or through the local distributor where one supplied the machine.

Where on-screen questions are the wrong answer

• Where the funder will not use the data. Questions that end up in no report only slow people down.
• Where the machine serves a tiny community. Even age band and postcode can identify someone; count dispenses and stop there.
• Where trust is already low. A new machine in a place with heavy policing may do better with no questions for its first months.
• Where outcomes need a real study. A machine survey is self-report from a self-selected group; it supports an evaluation but is not one.

Frequently asked questions

What questions should a harm reduction vending machine ask?

Only questions the funder will actually use, and as few as possible. The most common are whether this is a first supply or a refill, whether the last naloxone kit was used on someone, an age band, a postcode or area, and how the person heard about the machine. One to three questions per visit is a sensible budget, rotated so that different users answer different questions. Every question should be answered with a button, and every question needs a prefer-not-to-say option.

Can a harm reduction vending machine refuse supplies if someone does not answer?

It can, but for harm reduction supplies it should not. The guiding rule is ask, record, never block. NSW Health’s needle and syringe program guideline says participation in brief interventions, research and evaluation surveys must be based on the client’s informed and voluntary consent, and that services should avoid imposing interventions which may discourage future access. Victoria’s naloxone machine tender, as we summarised it, asked machines to confirm the recipient is 12 or over but not prevent access. Survey questions should never be an eligibility gate.

Should the questions be on the machine’s touchscreen or on the user’s phone?

Use the touchscreen for one or two quick questions at the point of release, because it needs no phone and catches every user. Use the phone for anything longer: registration, eligibility and follow-up surveys, because the person can answer in private and in their own time. Many programs combine the two: a single question on screen, plus a QR code to an optional survey. A screen-free machine can still collect answers if registration happens on a phone first, as on the HIVConnect and MyTest programs.

Is it safe to ask for a postcode or ZIP code?

It is low risk on its own and useful for showing a funder where users travel from. The risk rises when a postcode is combined with an exact age, gender and time of use in a small community, because together they can point to one person. Collect an age band rather than an age, report postcodes in aggregate, suppress small counts in published reports, and consider asking for a suburb or area rather than a full postcode where the population is small.

What should a harm reduction vending machine never collect?

Names, full dates of birth, phone numbers or email addresses typed into the machine, photographs, video or fingerprints, and free-text answers. None is needed to count supplies or evaluate a program, and each makes users less willing to come back. People who use drugs have told researchers they fear police access to biometric records. If a program needs a contact detail for a quota or follow-up, collect it on a separate registration page on the person’s own phone, never on the machine.

How many questions is too many?

More than about three, or anything that takes more than a few seconds at the machine. The Melbourne syringe machine study is the warning: three mandatory keypad fields before every order produced an estimated 50 percent invalid entries, with about a quarter of orders showing obvious repeated digits. People at a machine late at night want supplies, not a form. Ask fewer questions, rotate them, and accept that a smaller set of honest answers is worth more than a large set of junk.

How do the answers get into a funder’s report?

On KioskForce builds each release writes an order record with the items, quantities, machine, location and time, and the answers are stored alongside that record. The cloud dashboard can export the data by date range, and the program builds its report from it: units by product and site, first supply versus refill, reported naloxone use, age bands and areas, all in aggregate. Agree the report fields with the funder before launch, so the questions on the screen match what will be reported.

References

• NSW Health — “NSW Needle and Syringe Program”, Guideline GL2023_002 (18 January 2023). https://www1.health.nsw.gov.au/pds/ActivePDSDocuments/GL2023_002.pdf
• Harm Reduction Journal — “Analysis of four syringe dispensing machine point-of-access data 2017–2020 in Melbourne, Australia: machine utilisation and client demographics” (2022). https://pmc.ncbi.nlm.nih.gov/articles/PMC9768389/
• Russell E et al., Harm Reduction Journal — “A scoping review of implementation considerations for harm reduction vending machines” (2023). https://pmc.ncbi.nlm.nih.gov/articles/PMC10018614
• Marschke RM et al., Harm Reduction Journal — “Discreet but accessible: a qualitative study … optimal design of a harm reduction vending machine in rural Kentucky” (2025). https://pmc.ncbi.nlm.nih.gov/articles/PMC12117748
• Duplessy C, Reynaud EG, Harm Reduction Journal — “Long-term survey of a syringe-dispensing machine needle exchange program: answering public concerns” (2014), abstract via the Health Research Board library. https://www.drugsandalcohol.ie/22651/
• Victorian Department of Health — “Free naloxone machines rolling out across Victoria” (1 June 2026). https://www.health.vic.gov.au/news/free-naloxone-machines-across-victoria
• Office of the Australian Information Commissioner — “Read the Australian Privacy Principles” (accessed 11 October 2026). https://www.oaic.gov.au/privacy/australian-privacy-principles/read-the-australian-privacy-principles
• Information Commissioner’s Office — “Principle (c): Data minimisation” (accessed 11 October 2026). https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/data-protection-principles/a-guide-to-the-data-protection-principles/data-minimisation/

Related

• Harm reduction vending machines — the cluster hub.
• Controlled dispensing: limits, codes and cards
• Do harm reduction vending machines work?
• Naloxone dispensing machine specification
• How much a harm reduction vending machine costs
• Naloxone vending machines
• How free-vend health programs work

Talk to us about on-screen questions

Send the questions your funder needs answered, the languages, and whether users will register on a phone. We will come back with a screen flow, a machine and a price.