I am a clinical assistant professor at OU Health Sciences Center.  I am teaching an Advanced Diabetes Elective to 3rd year PharmD students who are interested in diabetes management and therapy.
I am developing a homework assignment re: insulin pump therapy.  I want to include exercises on how to arrive at starting basal rates, sensitivity factors, and insulin-to-carb ratios. 
I know that most pump trainers arrive at these starting settings in different ways. 
How do you arrive at these settings?
And what basics would you include when teaching students who are brand new to pump therapy concepts?
Are there any other concepts or objectives you would include for these students?
Thanks.
MH

A good place to find all the basics would be John Walsh's book Pumping Insulin.

While pumps get credit for improving BG control in general, they do so by virtue of forcing patients onto a Basal / Bolus regimen.  The only thing a pump can do better than Multiple Daily Injections (MDI) of insulin is provide the Basal dose.  Bolus doses from a pump are no different than bolus doses from an Insulin Pen or Syringe.  The pump's Basal dose, however, may not only be tailored hour by hour, unlike Glargine, Lantus, or even NPH, but it may be changed on-the-fly as needed for exercise or sports.

An extremely important issue that does not get enough attenton, unfortunately, is the need to establish the Basal dose experimentally by actually fasting, rather than by formula.  The formulas you will see tell you to take the Total Daily Dose and divide it in half, or perhaps take 40% of TDD as the daily Basal dose, then divide that by 24 for the hourly rate.

This is OK as a starting point, but the Dawn Phenomenon requires about twice this average rate in the early morning, while activity during the day calls for somewhat less (as little as half) in the afternoon and evening.  You can adjust the normal pump Basal rate by fasting 6 to 8 hours at a time, or go all the way by fasting from (after) dinner one day to (before) dinner the next.  After the dinner Bolus has balanced that meal, and you go to bed with normal BG, the pump basal rate should be adjusted to keep BG normal overnight and into the next day up to dinner again, checking BG every few hours.

Adjusting the Basal rate for sports or exercise, while the one place pumps outperform MDI, is not yet well enough understood for there to be accepted formulas and ratios that may be applied to the Basal rate.  Patients can start by cutting the rate in half, during and for a few hours after, and see how that works.  Heavier and longer activity not only require a greater reduction in the basal rate, but for a longer ime.  Unfortunately, no one can yet say how much for a given kind or degree of sports or exercise.

Otherwise, the Insulin Sensitivity factor (mg/dl reduction from 1 Unit), Insulin to Carb ratio (Units to balance 10g Carb), Insulin to Protein ratio (Units to balance 1 ounce Protein), and the BG rise from 10g Carb (absent Bolus Insulin), are all the same with a pump as with MDI.  While they are all very important for intensive insulin therapy, necessary for estimating both the Bolus itself and the various adjustments to the Bolus dose needed for good control (as detailed in my eTalk Strategies for progressively better Type 1 BG control), they are not unique to pumps.

David,
Thanks for your input to this topic.  Tell me more about the insulin to protein ratio.  I have used this with my patients but have never heard of any formal calculations used to determine what it should be.  I don't recall seeing it in the Pumping Insulin book.  What formula do you use?
Also, you mentioned the BG rise from 10g carb.  Tell me more about that.  I have not heard of that estimate or how it is used in pump therapy.
I will check out your e-talk.
MH

This is a bit off topic, because Bolus calculations apply to all Type 1 control, not just pump, but just over half (58%) of protein becomes carb over about 6+ hours, peaking at 2-3 hours.  Four oz of Turkey contains 24g protein, Steak 35g.  A nice list can be found here.  So 8 oz of Steak with 70g protein is equivalent to 40g carb, but spread out over a good 7 hours.  I need to adjust both Bolus and Basal insulin (bedtime NPH in my case since I don't pump) when I eat such a large steak dinner.  A much smaller amount, such as in a cheese or ham sandwich, isn't worth worrying about, because uncertainty in the effect of the bread has more impact on BG than this relatively small amount of protein.

The BG rise from 10g carb equals the product of the Carb Ratio (Units to balance 10g carb) and the Insulin Sensitivity (mg/dl reduction from 1 Unit).  My Carb Ratio is 1U/10g, Insulin Sensitivity is 33 mg/dl per Unit, so BG rises 33 mg/dl from 10g carb.  Your mileage will no doubt vary.

The BG rise from 10g carb is useful when on-board insulin calculations based on post-prandial BG, measured 90 minutes to 2 hours after eating, predict BG will fall below 60-70 mg/dl.  Say, for example, BG is 90 mg/dl an hour and a half after eating, and is predicted to fall to 60 mg/dl within another 30 minutes, and an ultimate 30 mg/dl after two more hours when the Bolus is used up.

You will want to eat enough carb to raise that 30 mg/dl to near 100 mg/dl, or by almost 70 mg/dl.  With a BG rise of 33 mg/dl from 10g carb, I'd eat at least 20 g carb within 20-30 minutes (probably sooner, but I wouldn't need to rush) to avoid hypoglycemia and wind up normal at the next meal/bedtime.

Rebound hyperglycemia is a real problem with hypoglycemia, because actually going low stimulates counter-regulatory action by the liver, and patients often eat far too much for their condition, taking 40 to 60 grams carb or even more when 10 or 20 would do.

But, again, this isn't strictly a pumping insulin topic.  It applies to all intensive Type 1 BG control.

Micki:
I do think that pumps can do more than just add variable basal rates for different times of day.  Variable basal rates are good.  Scheiner did a study showing that the average type 1 patient (322 in study) needed 5.2 basal rates per day.  It is really the only way to address dawn phenomenon with basal insulin.
With pumps you can also alter the bolus delivery, which you really can't with injections.  I have found extended boluses such as dual wave boluses to be very helpful for high fat and high protein meals.  Extended boluses are also helpful with gastroparesis in getting a better (although not perfect) match of insulin to glucose excursions.  Even if you are taking insulin for large protein consumption (over 5 oz), it works best to use a dual wave bolus since the protein effect on glucose won't happen in the first two hours.

Micki:  I didn't address your original question very well.  Bruce Bode MD has a booklet published by Medtronic that gives his preferences for starting the pump based on weight and pre-pump doses, as well as calculating carb ratios and insulin sensitivity factor.  These are starting points, and I agree that they need to be fine-tuned.  A lot of times providers simply "verify" the overnight basals and don't pay much attention to the daytime rates.  Rather than have patients fast for an entire day, I have seen good results using fasting for a portion of the day (such as skipping breakfast and testing BG every hour or two during the morning) to assess basal insulin rates.  I would recommend doing this on three different days rather than basing any changes on one day's data.  It takes a while to verify the rates this way, but it beats going without food for most of a day (at least in the minds of most of my patients).  Verifying the bolus doses is similar to how it is done on injections:  test pre meal and two hour postprandial and adjust as needed.  It is harder to verify the ISF since it may take three or four hours to see the full effect of a correction dose, and meals and activity during the day tend to interfere with this evaluation.

When testing basal rates, I usually have patients skip one meal and check BG every hour until the beginning of the next meal or I have them test BG every 2 hours overnight from HS to breakfast.  I have modified this for patients who would rather get the testing done quickly so that they will skip two meals or up to 12-14 hours of checking BG without eating or bolusing.  However, I am skeptical of having a patient (or myself) skip an entire 24 hours worth of eating to test basal rates.  Wouldn't fasting for that long of a period of time alter the results?  It seems that gluconeogenesis would skew the results if you were fasting for an entire day.
MH

Fasting a full 24 hours is a bit much, and the most I've done is from after one dinner to before the next, which is 20+ hours without glucose from the gut.  MiniMed published The Insulin Pump Therapy Book: insights from the experts in 1995, with a chapter by Bruce Bode, Establishing and Verifying Basal Rates, laying out how to estimate the hour-by-hour basal rate for one meal period (or overnight) at a time.  Once done, I still like the idea of going from one dinner to the next to verify.

You'd have to ask a physiologist to be sure, but I think gluconeogenesis is stimulated by low glucose (through the growth hormone pathway that is responsible for the Dawn Phenomenon), so maintaining normal glucose, even for a day, wouldn't skew the basal rate.  Eating, or not doing whatever is done on a normal day, is what confounds verifying the true basal rate, and I think Robyn is right about providers not paying attention to the basal rate during the day.  For me, the overnight rate is three times the afternoon rate.  Not getting the basal dose right for the entire day (while fasting) wastes the one place a pump out-performs MDI.