Life in Broome is sweet, a bit too sweet.

Broome is one of the centres of Aboriginal Health care in Australia.  One of the biggest disease burdens in the Aboriginal population is type 2 diabetes.  The stats are scary, and the complications of this potentially managable disease are rife in our community.  From preventable blindness to end-stage diabetic nephropathy – this is a crippling disease.  I see young people (younger than me) with these complications most days.  As hospital Docs we tend to see the pointy end of the problems – many years after intervention might have been useful, and it is easy to get a bit pessimistic about it all.

On the ward / ED we see crazy high sugars  [apologies to US readers – we use mmol/l – so multiply by 18 to get mg%]  I see ‘incidental’ BSLs of 25 [450mg] everyday, in entirely asymptomatic people!  I think many of us are chronically frustrated by seeing such remarkably uncontrolled DM and knowing that it is nigh on  impossible to ‘fix’ this problem in a 3 day admission – these folks need great primary care and education – NOT intensive inpatient care.  However, when we see a patient with a diabetic foot infection, pyelo or pneumonia we know that getting their BSL under control can help and get them better sooner – so we gotta do something to get the BSL down – but what?

We did an audit in Broome in 2010 and found our average inpatient BSL was 12.7 [~130]!!  And despite patients being on a QID regime – only 1.7 doses/day were completed on average!  The HbA1Cs were off the chart also – overall we sucked at glycemic control.  Why? – well bad diabetes, a high flux and mobile ward with frequently AWOL patients (sound familiar to anyone?).

So in order to relieve my frustrations I have become somewhat maniacal about trying to control inpatient hyperglycemia.  This area of the literature has been very active in the last few years with trials showing the benefits of tight glucose control in ICU, then trials showing that too tight control actually increased mortality – so my reading of this is that there is a benefit, but a tipping point where you might cause harm if you try to achieve perfect control.  The upshot – we aim for good glycemic control – single figures (BSL < 10 is nice, not hard to achieve and has a lower rate of hypos).

Traditionally we have used sliding-scale insulin QID regimes (SSI) on inpatients – this was one of the tortures of internship – adjusting SSIs almost at random to appease the nurses / sugar Gods.  I always suspected SSI was a bit nonsensical – like playing catch-up, and in 2007 the RABBIT-2 study showed superiority of a basal-bolus insulin regime (BBI).  For me this was killing multiple birds with one strategy.

  • BBI regime relies on a long-acting insulin – so missed doses are less significant
  • BBI makes sense-you anticipate the insulin requirement rather than react to hyperglycemia
  • Doses can be titrated in a simple manner to match the scenario/patient.
  • A more consistent ‘level’ of insulin should avoid the peaks and troughs of intermittent SSI boluses.
  • BBI is easier for the nursing staff – each dose requires asingle trip to the bedside, rather than 2 – this means less work, easier to time doses with meals and ‘catch’ patients when available.
  • You do need to counsel the patient with something like : “the needles are to get you better quicker, you don’t have to stay on insulin once you are better and ready to go home.”
  • One tip – always chart 24 hrs worth of insulin, starting with the lunch dose, and ending with breakfast dose to allow yourself time to review on the morning rounds.

I have added a permanent link to the Basal Bolus Insulin Order chart that we have trialled in Broome Hospital – the monitoring and doses are on the front with instructions on the back – it is pretty self explanatory.

Let me know – can you do this in your hospital?

We are planning a follow-up audit this year so will hopefully see some improved data!    Casey



  1. This just shows how difficult it is to make recommendations that are applicable to all settings. We obviously don’t have the same problems, so how could the solution possibly be the same? Why are only 1,7 out of 4 doses administered? Why does the nurse have to go to the patient twice with your former system? Where I work, in Sweden, anyone admitted with hyperglycemia (and all patients on insulin) will have their glucose levels checked 5-6 times daily and the nurse will administer insulin according to my orders (eg 4U if glu>10, 6U if>12, 8U if>15, 10U if>20). If the patient has underlying diabetes they will be started on a long-acting agent, too. It is really no big deal and definitely not during hospital admission. Your work with non-compliant patients poses totally different issues to your practice, than I face here. I imagine that’s both challenging and interesting.

  2. Graeme Pickford says

    Hi Casey,
    I was hoping to link this page in a newsletter (educational), but it doesn’t seem to load properly on two seperate connections I’ve tried (have to scroll and highlight to load text)
    But to also pose a question: What is “a diabetic diet” in your hospital? In Bathurst, it means a fixed portion (15g x carbs per serve) diet, BUT, this can vary between one and seven serves per meal (ie can vary between 15g and 105g of carbs per meal). If the docs are titrating bolus insulin to a seven fold variation in glycaemic load, it’s little wonder we mess it up. It would be more sensible to order a “fixed carb meal” (in Bathurst this means 45g carb per meal–still too high for my liking, but at least not a moving target)
    (If I ever finish the newsletter I’ll flick you a copy–have linked shamelessly to you and others in the previous one)

    • Thanks for the question Graeme
      Agree it is really tough to titrate insulin with random carbos going into the mix.
      Big problem we have is the patients are highly mobile, not nice, stay in bed and get a DVT types!
      They tend to leave the ward for long periods, meet with family etc on the grounds and eat who knows what! Probably why the previous audit showed only 1.7 BSL being done a day.

      The concept with BBI is to lean heavily on the longer acting insulin so there is more tolerance for random sugar intake etc. Not ideal but more practical and gets a better result than trying to use short-acting insulin aid.

      I fear we are looking at 2 very different populations?

      The link should be:


Speak Your Mind