Fine-tuning, or maybe completely revamping, settings

Curious for those on pump and CGM, how long did it take you to fine-tune your settings?

I'm more than three weeks in to using the Dexcom and am still trying to get my overnight basal rates sorted out. My basal rates have changed A LOT in the past three weeks and my blood sugar is slowly improving, but even now it's still varying by about 2.5 mmol/L (45 mg/dl) overnight. This is less variation than I was getting before, but my basal rates have changed drastically, so I'm really surprised that there hasn't been a bigger impact. I'm hoping that over the next week I'll be able to get it right, just in time for me to send this trail Dexcom back (so lucky the rep let me try a second sensor!) and buy my own a couple of weeks after that.

Also, I started the last sensor right around the start of my hormonal cycle, and it's been really interesting to see just how much it impacts my control. My hormones "kicked in" just over a week ago, and my average blood sugar shoes it. Excluding the first half week of sensor use, my average three weeks ago was 6.8 (122) with 70% of readings within range. My average two weeks ago was 7.8 (140) with 54% of readings within range. My average this past week was 8.5 (153) with 42% of readings within range. This is despite increasing my basal rates and ratios throughout the past week or so - basals have gone up by 3.6 units, I:C ratio has gone from 1:9 to 1:6, and ISF has gone from 2.0 (36) to 1.5 (27) - and still the big increase in average and decrease in readings in range. As I get my basal rates and other settings worked out, I'm hoping that I'll be able to be more aggressive with increaing my settings in the future (though part of the problem is the first few days of "hormones" is always a mixture of highs and lows, which makes it hard to increase settings until I'm consistingly high...). I'm not sure this trial sensor will last long enough for me to see the huge crash back down that should happen in about 10 days or so...

The other thing I have learned so far is that I really do have to change infusion sets every 24 hours. I was noticing that every second day I was runnig high, and I would start going low within hours of changing sets (because usually I changed sites and did a correction). Before, I figured if my site didn't feel super irritated it was probably okay, but even the "normal" irritation I get seems to degrade things.

Once I get my overnight basal rate sorted out I'll move on to fasting throughout the day to work on those. If this overnight one is any indication, they may each a month! After that I'll have to test my carb and correction ratios.

I get my next A1c in about a month and will be very curious what it will be - and even more curiosu what I might be able to achieve six months from now.

Jen - I've found that my basal rates are never done, as in "dialed-in" and finished. When I get good results, like a flatline overnight, it may go on for a few nights but then seems to degrade in some way. It may start to trend higher or lower. Sometimes I think I over-react to these changes. It's better to wait for a few days to see if it becomes a pattern. I define a pattern as a trace that repeats itself three times or three days in a row.

I've experienced a slight rise recently in overnight BGs but not out of range. So I adjusted the basal rates by 0.1 unit/hour for the period two hours before. Last night I went low, into the high 50s (3-3.3 mmol/L) during that time. So now I've removed that recent change and we'll see.

What I've come to learn is that good diabetic control means responding almost continuously to a changing BG environment. The measure of control that results depends on how quickly you adjust your counter-measures. It reminds me of an athlete in a sport like downhill skiing. The skier must make many adjustments, some small, some large, in order to ski the course well. Sometimes the adjustments occur at several each second and sometimes much less so.

I think the CGM permits valuable insights into our own body's response to various inputs. You will figure out your own puzzle but I don't think that puzzle remains static.

My overnight has been a pretty consistent pattern of my blood sugar first rising and then falling quite a bit. I've eliminated the rise, but it still falls after about 3 AM (which means I can't go to bed in range or I end up low). I don't expect overnight to be perfectly flat - I think that's asking too much! - but at least staying within range (which I have set to 70-180) would be good to aim for.

I've been talking texting daily with a friend of mine who has Type 1 and uses a Dexcom and has an A1c within the normal range, and she has provided some great suggestiosn. Like you, she said she is always adjusting her basal rates and never takes the same basal TDD each day, but she does have a pattern "dialled in" and then uses a lot of temporary basal rates to adjust for what she is doing. When I talk about pump settings I don't mean that they will never change. But I do think there are basic settings that can be dialled in, and then adjusted for hormones/activity/etc. as needed every few days or so. Maybe this isn't correct, but there must be something to things like basal testing and ratio testing, otherwise no one would ever do them...

One of the great things about the Dexcom is that I've been a lot more responsive than I used to be. Previously I would often let mysel run high for an entire day before doing anything. Now, after 2-3 hours of out of range it starts to annoy me. The downside is that probably about 50% of the time treating highs or lows has sent me in the other direction, even with pretty conservative treatment (8 grams of carbs for a low, or 0.75 units over what my pump suggests for a correction of a high that's been there for 3-4 hours).

I agree with Terry that I'm never "done" as there always seems to be some drift or wobble. It's hard to say (particularly not keeping any records, other than in my head...horrible scientist!!), whether some of it might be due to my drifting food and beverage habits and how much is due to changes in exercise and how much of it might just be insulin needs varying from time to time.

Before I had a CGM, I would let bad trends ride for days, sometimes weeks. I think how quickly we respond to changes, to a large extent, determines our overall success.

I think successfully finding a good basal pattern means finding the right BG "neighborhood." From there it means making minor adjustments or "nudges" as AR likes to call them. When I get to the scenario where only nudges are needed, I feel I've arrived! By the way, I've even split a glucose tab in half so that I treat with 2 grams of carbs to correct a minor low. Insulin corrections of 0.1 or 0.2 units are a nudge in the other direction.

And yet I hear of people running +200% basal rates during highs to get them down more quickly. Maybe next time I will try 6g of carbs. Part of my complication is that when I dont' have water I can't swallow glucose tablets easily (due to my EE) so I've started using those gel things if I go low while I'm on the bus or whatever. Half of one of those is about 8g of carbs. I don't know if I could estimate more precisely than that, though maybe I could sort of try to guess at 1/4 which would be 3-4g of carbs.

One thing I am finding with highs is that often I will go hgih and be high for half a day without budging, and then a relatively minor correction (like 1 unit) will all of a sudden send me low. I'm not sure if this is a basal issue or what, but it's another thing I have to figure out.

I agree with you about ballpark. That is really my only goal. It's not realistic to stay in range all the time, although at the moment my range is 70-180 and I haven't had a day where I've stayed within those lines. If I could go all day staying within that wider range (since my *real* range is 70-125 as set by my endo), then I would be satisfied.

But there is one thing I'm confused about. You and AR (and maybe others) never do things like fasting for a basal test to get it "set" correctly??

I last did the fasting basal tests in 2012 when I drastically reduced my carb consumption. I did that with Gary Scheiner’s help. Since then I have made small but frequent changes to my basal pattern.

Using the CGM it’s easy to make educated changes to the overnight basal rates since you are fasting. I also like to do one intermittent fast of 24 hours each week. I skip my late breakfast and fast from dinner to dinner. That gives me insight into my morning and afternoon rates. I’ve been able to make changes to my evening rates without the benefit of a fast.

Starting from scratch is the way to go if you think your rates are not close. Once you arrive in the right neighborhood, small tweaks can be done without a fast. All of this assumes that you’re not making any large changes to your exercise or food patterns.

When I tried using the +200% basal rate to bring down a high BG, I ended up gaining about six pounds in a month. This was after an extended period of stable weight. Now I use an IM shot. It works quickly, well, and I don't gain weight.

Took about 2 months to get it 'dialed in'. But then it's always changing, a little - a lot when sick of course.

Interestingly enough, once I got it 'right' I was pretty much a flat-liner. I often made adjustments to keep it that way...until January when I was hospitalized with a post-op MRSA infection in my foot. My BG went crazy...mostly low. I made about 30 temp basal rate changes in the 8 days I was in the hospital. I continued to run low after getting out, requiring constant adjustments between 75% - 110% (of my old normal). Then one day, all of a sudden my BG doubled! Went from a 85% to 180% basal rate in one day, just to keep it down to about 120!

Still struggling a little, the last several days I've needed 110%, then all of a sudden need a 80% to keep from crashing.

Hopefully it will all work out after everything is healed and I'm off the IV anntibiotics (another 3 weeks or so). Hope to be a flat-liner again soon.

Hi Jen,

I'm still in the process of fine tuning things since I've only been on the pump about 2 months. I haven't made any basal changes myself yet, just temp rates when I need them. Last endo appointment last week, I saw my cde also and discussed some things. She looked at my pump & cgm data which I finally uploaded to dexcom. We changed some of the basal rates due to my bg. Dexcom data says I have no apparent patterns, which kind of surprised me, and I wonder if that is true. So she said I have to play it by ear and adjust the basal according to what is going on with bg- like maybe go to sleep with a temp rate if I'm higher. I'm not sure what I will do yet, so far I'm still up monitoring things unfortunately.

When I was training they had me set to do 4 different basal tests, I had to cancel the first one due to high bg due to too little insulin for the meal, then I did a 23 hour one and they said to stop after that. I don't recommend the 23 hour test because it can cause your liver to react, two 12 hour tests would be better. I may do more again at some point but since I developed ketones and vertigo after the 23 hour one I won't be doing that again.

Can Terry or others share what sort of basal rates they are talking about when they talk about having variations from one night to the next, and the kinds of tweaks that are made? It would be nice to have something to compare to and see what I should be aiming for. Last night I varied by 5 mmol/L (90 mg/dl), the night before that I varied by 7 mmol/L (126 mg/dl), the night before that by 2.5 mmol/L (45 mg/dl), the night before that by 3 mmol/L (55 mg/dl), the night before that by 6 mmol/L (108 mg/dl), and the night before that by 6.5 mmol/L (117 mg/dl). Somehow I think this is not the type of variation people are talking about when they talk about making tweaks. I get the impression tweaks are made if people usually wake up at 5.5 (100) and then suddenly start waking up at 8.0 (145) or something like that, and I am not even close to that. Maybe I am wrong, though.

Dexcom doesn't give any patterns for me, either, when looking at 7, 14, and 30 days of data. For 14 days it does have one thing that says the most significant pattern of highs is between 8:00 AM and 9:00 AM, but it only says (1 high found) so I wonder if this is only because one day a week and a half ago or so I forgot to bolus and spiked to 19.3, which is the highest I have reached on the Dexcom so far. Otherwise, it says no significant patterns found for everything.

I'm looking so farward to my next endo appointment in March. In the past he has never been able to spot any real patterns, but I feel like the Dexcom does reveal more patterns than testing 10x a day did, so I'll be very curious about his thoughts on everything.

Jen, my overnight variation in BGs is much less than yours. Here's a recent trace that I wanted to change:

I decreased my 11:00 p.m. and midnight to 2:00 a.m. basal rates from 0.6 units/hour to 0.5 units per hour. That led to this result:

You are dealing with much larger BG variability overnight, therefore your tactics need to be coarser. Reducing BG variability should be your first goal. The nature of wider variability requires a higher BG average overnight to remain safe. So I wouldn't worry about the level of your overnight BGs so much as the amount of swing you experience.

The best way to test this is to go to sleep without any active nutritional insulin on board as well as no active food. In other words, no food or snacks in the last three or four hours. If you go low, of course, treat it but then put off the test until another night. Then change your basal rates two hours in advance of any rolling highs or lows. Increase rates if going high and decrease rates if going low.

Your goal is to flatten that overnight line so that ideally it doesn't vary by more than about 2 mmol/L (36 mg/dl). Once you have a relatively flat line then you may address the overall level of that line. Let's say that line averages 7 mmol/L (126 mg/dl). Then you just need to increase all the rates starting at two hours before sleep by a significant step. I like to use 0.1 unit steps as the smallest incremental change. I find thinking in terms of 0.025 steps confusing. My basal rates range from 0.4 units/hour to 1.1 units/hour. If your rates are much smaller, like 0.05 to 0.3 units/hours then a smaller increment, like 0.05 might make sense.

To summarize, reduce overnight BG variability first, then reduce the average BG of the overnight level. Make changes two hours before the target time frame. It's best to wait for a three time/or day pattern to emerge before making changes.

I forget what your style of eating is. My tactics would not work as well with someone that eats a lot of carbs. I had pretty good success eating less than 70 carbs per day and much better results now that I target 30 grams of carbs per day. I think the inherent errors built into the consistency of insulin action and digestion make taking large doses of insulin a game of chance. As I've said many times, fewer carbs = less insulin = smaller errors.

Your inquisitive nature will serve you well with this new visibility of data that the Dex enables. Don't give up. You are on the right path.

That is odd, I would have thought the point of that program is to find patterns so we can adjust things more easily maybe than having to pour over data for hours finding a pattern? I'm going to ask my CDE why it doesn't find the patterns, I assumed it was because mine vary too much? One high doesn't sound like a pattern. She did find some patterns though looking at the data and pointed them out- like I often rise in the afternoon, but sometimes I go hypo too with exactly the same food etc. so I'm not sure what to do.

I don't know when my worst highs are but lately I have been spiking up around 4 am consistently for 2-3 days but still it is either high or low for me, never totally consistent or predictable. I went to 209 Tues 4am for no reason apparently. I thought maybe it was my inset but it came down fairly rapidly with a bolus and temp rate so it wasn't.

I hope your endo gives you some good help/input!

The patterns you're looking for are the %age of numbers in range. Medtronic has these piechart reports that show like "after breakfast" with low/inrange/high by %age in a pie chart and it breaks down the whole day. So I look at that chart and go "hmmm, which time is the most jacked up?" (which is often breakfast, of course...) and say like post lunch has spiked the lows to 40% of the time. I figure either 1) I'm not counting carbs right, always a possibility, or my lunch bolus ratio needs to be turned up (as up is down with those...) to shave off some insulin. I'd probably only nudge it .1U/ hour for a post-postprandial number. If it's before lunch that's off, I'd consider changing basal. I don't do formal "clear" tests like Terry recommends as I am sort of a creature of habit, come home, work out (ok, I blew it off today...) and then eat and sit around. I don't eat huge in the evenings but sometimes will split a bag of popcorn 3 ways, booze it up, etc. If my overnights are out of whack, it's usually pretty perceptible and I'll try to see what time the "woah" starts to occur and head it off at the pass.

Although I am not a low-carber, I would still go along with Terry's suggestion of going to bed "clear" with no bolus-on-board. AR's on-the-fly approach works well if your basals are already well set and you just need to tweak, but if they are significantly out-of-whack it's better to cut out the "noise".

I rarely adjust any of my basal segments by more than 0.05 u/hr. Even this much will typically result in a change of around 3 mmol/L in my subsequent levels. Usually changing by 0.1 will have too big an effect.

Terry: I would have considered 70g carbs/day as low carb. I cannot imagine how you can eat only 30. Things like vegetables that are highly nutritious still contain small but significant amounts of carbs. I reckon that my diet contains >30 g per day just in what I call mickey-mouse carbs (half a decent size onion is around 8g).

Jen: I wouldn't get too fixated on your upcoming HbA1c. Running your glycaemic control on the basis of HbA1c is a bit like steering a car by looking through the rear-view mirror. The Dexcom is accurate enough to give you real time views of where you are going (and where you should be going). FWIW my A1c went up slightly when I started CGM, largely because I was removing the averaging effect of being frequently hypo.


My Dexcom reports don't show anything in the Pattern Insight chart either. It says there are no high or low readings, but there definitely are. Might it be a glitch in the software?

SD - I think the confusion about the Studio program's conclusion hinge on the definition of a "pattern." It usually means more than one occurrence for a few days or instances in a row. So, a single high or low excursion does not trigger a pattern report.

Significant doesn't apply to single number. It is showing a pattern of high/low numbers at the same time, for several days.

The Patterns screen is useful when your routine is somewhat constant - meals at same time, sleep/wake at same time, etc.
It is looking for patterns within the same hours across the number of days that are displayed. If you have high number of days, it is less likely you will see patterns, because they average out.

If you look at the bottom portion of the screen, there is a representation of highs/lows (hills and valleys). The more frequent and higher/lower the number is within each hour time slot, more shading will be shown above/below the line (which represents your target range).
Once the hill/valley gets to a certain height, it turns color, and shows as a pattern on top right.

Try setting your target at different ranges, and the number of days, and it might make more sense.

If your meal times are generally 6,12,6, and you regularly go high after meals, that might show a pattern. But if your meal times vary, and you go high, it would be spread out, and not show as a pattern.
I think it works best using 7-14 days.

I find it most useful for overnight basal, since my mealtime are quite varied, and I get immediate feedback when I miss the mark on my meal boluses.

I was on the phone with Dexcom today- to let everyone know I really didn't get a good explanation of why no patterns are showing etc.. they talked me through setting up Portrait at Safari, which is a web based app that Mac users can use to upload their data(use Safari if you do this I was told). Studio is being eliminated and Portrait is going to take over.

Why they didn't design software platforms from the getgo for everyone I don't know- it's plain stupid imo. Plus I was told by someone in August that they would have something for Mac users in a few months. Now it looks like Portrait is the only thing that is going to happen. One tech told me to switch the setting to 30 days in the patterns data area( apparently this is the only data currently available at Portrait now which is being worked on he said) but it made no difference, there were still no patterns.

I really just don't see the point of this software if it doesn't give you that. And even *more* I don't see the point of making this hard to access for anyone, ie: mac and linux users.

My cde printed out 4 pages of various data, one which did show averages/median numbers etc. but there is no way you can look at the graph of 30 days bg dots in the patterns/trends data and figure anything out period, it looks like a tangled web or yarn, lol.

I'm going to go to diasend and see if I see more there which is where I uploaded my data to at first for my cde.

Another thing I was unable to do due to using macs is to upgrade the software for the dex itself, my cde had to do that for me on her computer, which is ridiculous imo.

oh, btw, another person I spoke with, the mac expert, told me that they have to get everything approved by the fda including the software, kind of as an excuse for why there was nothing for mac users, but I don't see that as a valid excuse because tandem has software platforms for everyone- I have no problem uploading my data, looking at it myself and sharing it with my cde.

I guess it could be a glitch??? I don’t know.