Insulin and Support Workers: What Is Allowed and What Goes Wrong

Insulin and support workers: what is allowed, what goes wrong, and how to check a provider

Insulin is one of the highest-risk medications in community care. It has no safe margin above the correct dose, it is supplied in look-alike devices, it is given several times a day by people who are not clinicians, across shift handovers, in houses where the roster changes weekly.

And it appears in the NDIS high intensity skills descriptors, where subcutaneous injections are listed as a distinct area requiring demonstrated worker competency — the descriptor itself naming insulin as the most common example.

So a provider delivering insulin support has specific obligations, and a family placing a person in that provider’s care has specific questions they are entitled to ask. This guide is about both, plus the thing that matters most: the particular ways this goes wrong.

We support NDIS participants across New South Wales, including people who need insulin administered under clinical plans written by their own treating teams. We do not write those plans and nothing here is medical advice or an instruction to perform any task. Every clinical detail belongs to the person’s own diabetes management plan, which governs.

The rule that everything else hangs off

A support worker never decides an insulin dose.

Not a bit. Not in an emergency. Not because the reading looks high and the worker has seen this before. Not because the person asked them to. Not because the plan is ambiguous and somebody has to make a call.

Dose decisions are prescribing decisions, and they belong to the prescriber. The worker’s job is to administer what has been prescribed, exactly as written, or to escalate.

This sounds obvious and it is the rule most often bent, usually with good intentions. The bending happens in grey areas, and the grey areas are worth naming precisely.

Grey area one: correction or sliding-scale doses. A plan says: if blood glucose is above 15, give an additional 4 units. Is a worker following an instruction, or making a clinical judgement?

They are following an instruction — provided the instruction is complete. A usable protocol specifies the exact threshold, the exact dose, the exact insulin, the maximum number of corrections in a period, what to do if the reading is above a second higher threshold, and when to stop and ring. It is written and signed by the prescriber, it sits with the medication chart, and it leaves nothing to interpretation.

An instruction that says “give extra insulin if high” is not a protocol. It is a delegation of prescribing to an unqualified person, and a worker who follows it is being set up.

Grey area two: carbohydrate counting. For somebody on a flexible insulin regimen, the dose depends on what they are about to eat. That is a genuine calculation, and whether a worker can do it depends entirely on whether there is a documented ratio from the prescriber, training specific to that person, and demonstrated competency. Where that exists it is administration. Where it does not, it is guessing with a drug that has no safety margin.

Grey area three: withholding. A worker who decides not to give a dose has made a clinical decision just as surely as one who gives extra. If the person has not eaten, is vomiting, is unwell, or has a reading that worries the worker, the plan must say what happens — and if it does not, the answer is to ring, not to decide.

The three tiers of task

Worth separating clearly, because they attract different requirements.

Ordinary support, not high intensity. Prompting and reminding. Helping somebody get their equipment out. Supporting them to do their own finger prick. Preparing meals. Encouraging activity. Recording what was eaten. Observing and reporting.

Blood glucose monitoring sits here or close to it in most arrangements — it is not one of the eight high intensity descriptors — but it still needs training, because the technique affects the result, and it needs a written protocol for what readings trigger what response.

High intensity support. Administering the insulin injection. This is a subcutaneous injection under the NDIS high intensity skills descriptors, and it requires the provider to demonstrate that the specific worker is competent to perform the specific task for the specific person, with clinical oversight and a current plan from the treating team.

Never a worker’s task. Deciding or changing a dose. Changing the insulin type. Adjusting pump basal rates or ratios. Interpreting an HbA1c. Deciding a medication is no longer needed. Diagnosing anything. Managing a diabetic ketoacidosis presentation beyond recognising it and calling 000.

What a diabetes management plan has to contain

If a provider is administering insulin on the basis of anything less than this, there is a problem. The plan should come from the person’s treating team — GP, endocrinologist, credentialled diabetes educator — and should be current, not three years old.

• The diagnosis and type, and whether the person is insulin-dependent
• Each insulin by exact name and device, the dose in units, the time, and the relationship to food
• Any correction protocol, complete and unambiguous as described above
• Blood glucose monitoring: when, and the target range for this person
• What readings trigger what action, with specific numbers — the low threshold, the high threshold, the escalate-now threshold
• The hypoglycaemia protocol, in full, including what this person’s low actually looks like
• When to test for ketones, and what to do with the result
• The sick day plan — what changes when the person is unwell, because this is when people die
• Who to ring, in order, with after-hours numbers
• What to do about a missed dose
• Injection sites and the rotation plan
• Dietary information relevant to the regimen
• Review date

A plan without a sick day section and without a complete hypoglycaemia protocol is not a usable plan, and those are the two most commonly missing parts.

The error modes, which is the useful part

This is where experience is worth more than policy. These are the specific ways insulin administration goes wrong in supported accommodation.

Wrong insulin. The single most dangerous error. Many people are on two insulins — a rapid-acting one with meals and a long-acting one once daily. The pens look similar, are stored together, and have names that are easy to confuse under time pressure. Giving a long-acting dose as though it were rapid-acting, or giving a rapid-acting dose in place of the nightly long-acting one, can cause severe prolonged hypoglycaemia.

What prevents it: physically separating the two insulins, labelling them in a way that cannot be misread, never storing them in the same container, and a chart that names the insulin and the device at the point of administration rather than in a folder in the office.

Units versus millilitres. Insulin is dosed in units. It must only ever be drawn with an insulin syringe or delivered by an insulin pen. Drawing insulin with an ordinary syringe graduated in millilitres is a classic route to a tenfold overdose, and it is a known cause of death. If a worker is reaching for a standard syringe to give insulin, something has gone badly wrong upstream.

Double dosing across a handover. The outgoing worker gives the dose and does not sign. The incoming worker sees an unsigned chart and gives it again. This is one of the most common serious medication errors in community care and it is a documentation problem, not a competence problem.

What prevents it: signing at the point of administration, never later; a handover that explicitly states what was given and when; and a chart that lives where the medication is.

Missed dose, not escalated. A worker arrives, the dose was due two hours ago, nobody gave it, and the worker does not know whether to give it now. The plan has to answer this, and if it does not, the answer is to ring.

Wrong timing relative to food. Rapid-acting insulin given and then the meal delayed, or not eaten, is a direct route to hypoglycaemia. Rapid-acting insulin given well after the meal does not do its job. The plan should state the relationship and the worker should not give a mealtime dose until the food is actually in front of the person.

Injecting into lipohypertrophy. Repeated injection into the same site builds up firm, rubbery tissue under the skin. Insulin absorbed from it is erratic — sometimes under-absorbed, giving unexplained high readings, and sometimes suddenly released, giving an unexplained low. It is extremely common, it is invisible unless somebody feels for it, and it is a major cause of unstable glucose that gets blamed on diet.

What prevents it: a documented site rotation plan, actually followed, and somebody periodically feeling the sites. A worker noticing a firm lump should report it rather than injecting around it.

Needle reuse. Needles are single-use. Reused needles blunt, hurt, and contribute to lipohypertrophy.

Storage failures. Unopened insulin belongs in the fridge, not the freezer — frozen insulin is destroyed and must be discarded. An in-use pen is generally kept at room temperature and has a limited life after first use, commonly around 28 days depending on the product. Insulin left in a car in an Australian summer is no longer reliable insulin.

What prevents it: dating the pen when it is first used, a fridge thermometer, and never transporting insulin in a hot vehicle without a cooler.

Air shot omitted. Pens generally require a priming dose before each injection to clear air from the needle. Skipped, the delivered dose is less than the dial shows.

Not holding the pen in long enough. Pens require the button to be held for several seconds after the dial returns to zero. Withdrawn early, part of the dose runs back out. This is a very common technique failure and it produces unexplained high readings.

Hand washing before a finger prick. Sugar or food residue on a finger produces a falsely high reading. A falsely high reading can produce a correction dose that the person did not need, and then a hypo. Washing and drying hands is not a hygiene nicety here — it affects the dose.

Reading the wrong number. Continuous glucose monitors display a current estimate and a trend arrow, and the two together mean something different from the number alone. A reading of 5.5 falling fast is a different situation from 5.5 steady. Workers need to be taught to read the arrow, not just the figure.

Insulin pumps, briefly and firmly

An insulin pump delivers a continuous background infusion plus mealtime doses, and it is programmed with settings that only the prescriber and the person’s diabetes team should change.

A support worker should generally not be adjusting basal rates, insulin-to-carbohydrate ratios or correction factors. What they may be trained to do, where it is in the plan and competency is demonstrated, is assist with a mealtime bolus the pump is programmed to deliver, respond to alarms, and recognise pump failure.

Pump failure is the thing to plan for. Because a pump uses only rapid-acting insulin, an interruption — occlusion, dislodged cannula, empty reservoir, battery, device failure — means insulin delivery stops entirely, and someone on a pump can move toward ketoacidosis within hours rather than days. Every house supporting a pump user needs to know where the backup insulin pens and the backup plan are, and what the escalation trigger is. If nobody can answer that, it is the first thing to fix.

Documentation that is worth something

• A medication chart that lives where the medication is, not in the office.
• Signed at the point of administration. Every time, by the person who gave it.
• Blood glucose readings recorded with the time, every time, including the ones that look fine.
• What was eaten, where the regimen depends on it.
• Every hypo recorded — time, reading, symptoms, what was given, how long to recover, and the suspected cause. A pattern across six weeks is what gets a dose changed. Six scattered entries nobody reads are not.
• Every error recorded and reported, including the near misses.
• Site rotation recorded.

The reason to be strict about this: diabetes dose adjustment is driven by patterns. An endocrinologist looking at six weeks of properly recorded readings, meals and hypos can fix a regimen. The same clinician looking at a sparse chart with gaps and no times will change nothing, and the instability continues.

What a medication error requires

Insulin errors happen, including in good services. What distinguishes a safe provider is what happens next.

• Act on the clinical risk first. Follow the plan, ring the number in the plan, or ring 000 if the person is unwell. An insulin overdose is a medical emergency regardless of how embarrassing it is.
• Do not wait to see. With insulin, waiting is the wrong instinct.
• Record it factually, with times and the actual dose given.
• Report it internally under the provider’s incident process.
• Consider NDIS reportable incident obligations. A serious medication error causing harm may be a reportable incident to the NDIS Quality and Safeguards Commission, which is on 1800 035 544.
• Review the cause, not the person. Most insulin errors are system errors wearing a worker’s name — two insulins stored together, a chart in the wrong place, a handover that lost information, a plan that was ambiguous.

A provider whose response to an insulin error is to performance-manage the worker and change nothing about the system will have the same error again.

Competency: what real sign-off looks like

This is where providers differ most, and where the words all sound the same.

Not competency: a twenty-minute online module. A generic diabetes awareness certificate. An induction. “The team has all been trained.” Somebody showing a new worker once during a shift.

Competency, in the sense the NDIS high intensity descriptors intend:

• Theory training on diabetes, insulin and hypoglycaemia
• Person-specific training on this individual’s regimen, devices, plan and presentation
• Supervised practice
• Observed assessment of the actual task on the actual person, signed off by an appropriately qualified clinician
• A dated record of that sign-off, naming the worker, the task and the person
• Reassessment at intervals, and after any change to the regimen
• Clinical oversight available for questions

The point of person-specific assessment is that it is not a generic skill. A worker competent with one participant’s pen and plan is not automatically competent with another’s pump and different insulin.

The questions that test a provider

Ask these, and ask for documents rather than reassurance.

1. How many of the workers rostered to this person this month are individually signed off to administer their insulin? A good answer is a number and a document. A poor answer is “all our staff are trained.”
2. Who signed them off, what are their qualifications, and when?
3. Can I see the current diabetes management plan, and who wrote it?
4. Does it include a complete hypoglycaemia protocol and a sick day plan? If either is missing, that is the gap.
5. What is the escalation threshold, and can a worker ring an ambulance without asking a manager first? The answer must be yes.
6. Where is the medication chart kept? If the answer is the office rather than where the insulin is, expect errors.
7. How are the two insulins stored and distinguished?
8. What happens if a dose is missed, and who decides?
9. How often does a clinician review the chart and the readings?
10. What is the plan if the pump fails, and where is the backup insulin?
11. How many insulin errors have you had in the last twelve months, and what changed as a result? A provider that says none is either very small or not counting.
12. Who provides clinical oversight, and how do I contact them?

That eleventh question is the most revealing one. A provider that can describe an error and the system change that followed is safer than one that claims a perfect record.