1 00:00:00,000 --> 00:00:08,469 foreign 2 00:00:00,500 --> 00:00:08,469 [Music] 3 00:00:16,020 --> 00:00:20,880 they'll be doing a lovely talk on 4 00:00:19,440 --> 00:00:23,100 diabetes 5 00:00:20,880 --> 00:00:26,039 at heart keeping people with diabetes 6 00:00:23,100 --> 00:00:28,500 alive is all about data Lana has been 7 00:00:26,039 --> 00:00:30,180 diabetic for 20 years insulin using for 8 00:00:28,500 --> 00:00:33,120 a decade and now writes technical 9 00:00:30,180 --> 00:00:35,100 documentation About Time series data 10 00:00:33,120 --> 00:00:36,899 Steve has been an open source developer 11 00:00:35,100 --> 00:00:38,579 for 20 years as a and has an 12 00:00:36,899 --> 00:00:41,520 uncontrollable urge to solve every 13 00:00:38,579 --> 00:00:43,320 interesting problem he comes with 14 00:00:41,520 --> 00:00:45,120 together they have interesting dance 15 00:00:43,320 --> 00:00:46,620 problem that they set out to investigate 16 00:00:45,120 --> 00:00:49,020 today we get to hear what they 17 00:00:46,620 --> 00:00:51,320 discovered so please welcome Lana and 18 00:00:49,020 --> 00:00:51,320 Steve 19 00:00:54,840 --> 00:00:59,160 so we'd like to begin by acknowledging 20 00:00:57,059 --> 00:01:01,320 the individual owners of the land that 21 00:00:59,160 --> 00:01:03,660 we are giving his presentation on today 22 00:01:01,320 --> 00:01:05,339 and pay our respects to Elders past 23 00:01:03,660 --> 00:01:07,200 present and emerging 24 00:01:05,339 --> 00:01:09,060 we'd also like to acknowledge this land 25 00:01:07,200 --> 00:01:11,220 which we benefit from occupying was 26 00:01:09,060 --> 00:01:13,920 stolen and sovereignty was never seeded 27 00:01:11,220 --> 00:01:16,640 by the Warren G Roy roaring people of 28 00:01:13,920 --> 00:01:16,640 the cooler Nations 29 00:01:17,880 --> 00:01:24,360 one in 20 Australians have diabetes 30 00:01:21,540 --> 00:01:26,280 and in 2020 it was the seventh most 31 00:01:24,360 --> 00:01:29,159 common way to die 32 00:01:26,280 --> 00:01:31,259 isn't that depressing 33 00:01:29,159 --> 00:01:34,080 luckily this talk isn't really about 34 00:01:31,259 --> 00:01:35,820 diabetes that's for a few reasons one 35 00:01:34,080 --> 00:01:38,100 because this is a technical conference 36 00:01:35,820 --> 00:01:40,979 not a medical one and two because 37 00:01:38,100 --> 00:01:44,820 neither of us are doctors 38 00:01:40,979 --> 00:01:48,659 what this talk is about is data 39 00:01:44,820 --> 00:01:51,000 data is we think inherently interesting 40 00:01:48,659 --> 00:01:53,340 because if you only have enough of it 41 00:01:51,000 --> 00:01:55,140 what you have is information 42 00:01:53,340 --> 00:01:56,399 and we are lucky enough to be living in 43 00:01:55,140 --> 00:01:58,500 2023 44 00:01:56,399 --> 00:02:00,840 where we have these amazing mechanisms 45 00:01:58,500 --> 00:02:02,880 that can take in large quantities of raw 46 00:02:00,840 --> 00:02:05,280 data and turned it into really 47 00:02:02,880 --> 00:02:06,799 interesting pieces of information 48 00:02:05,280 --> 00:02:09,899 and when you have enough information 49 00:02:06,799 --> 00:02:11,879 especially when it's about yourself and 50 00:02:09,899 --> 00:02:14,099 the way your body works you get to make 51 00:02:11,879 --> 00:02:17,160 decisions that can in some cases make a 52 00:02:14,099 --> 00:02:20,099 real material difference to your life 53 00:02:17,160 --> 00:02:21,300 things like if I change this recipe in 54 00:02:20,099 --> 00:02:23,459 this way 55 00:02:21,300 --> 00:02:25,260 will it make it easier or harder for my 56 00:02:23,459 --> 00:02:27,000 body to process 57 00:02:25,260 --> 00:02:30,120 should I drink a coffee immediately 58 00:02:27,000 --> 00:02:31,200 before after working out or wait till 59 00:02:30,120 --> 00:02:33,420 afterwards 60 00:02:31,200 --> 00:02:35,580 or even things like what brand of pasta 61 00:02:33,420 --> 00:02:38,000 or what bread is better for me than this 62 00:02:35,580 --> 00:02:38,000 other bread 63 00:02:38,400 --> 00:02:43,920 so in order to talk about the data and 64 00:02:41,160 --> 00:02:45,060 information we can get out of it we kind 65 00:02:43,920 --> 00:02:48,360 of need to give you a little bit of a 66 00:02:45,060 --> 00:02:51,480 primer now we're both still not doctors 67 00:02:48,360 --> 00:02:54,120 so please bear in mind that this is our 68 00:02:51,480 --> 00:02:56,580 understanding of diabetes based on being 69 00:02:54,120 --> 00:02:58,140 in my case a person who lives with it or 70 00:02:56,580 --> 00:03:00,300 in Steve's case someone who lives with 71 00:02:58,140 --> 00:03:02,220 someone who lives with it 72 00:03:00,300 --> 00:03:04,080 uh so everyone's experience with chronic 73 00:03:02,220 --> 00:03:06,660 illness is different and what works for 74 00:03:04,080 --> 00:03:08,819 me won't necessarily work for you or 75 00:03:06,660 --> 00:03:11,640 your aunt or your friend or your cousin 76 00:03:08,819 --> 00:03:14,220 from Trinidad on Facebook in this talk 77 00:03:11,640 --> 00:03:17,099 we are going to touch on some mildly 78 00:03:14,220 --> 00:03:18,959 gross stuff like bodily fluids and we 79 00:03:17,099 --> 00:03:21,599 also do briefly mention Eating Disorders 80 00:03:18,959 --> 00:03:23,879 later on so now it's absolutely the time 81 00:03:21,599 --> 00:03:25,500 to head out if that is not your jam and 82 00:03:23,879 --> 00:03:28,580 there's absolutely no judgment if you 83 00:03:25,500 --> 00:03:28,580 choose to to head out 84 00:03:28,800 --> 00:03:33,300 so according to most things you read 85 00:03:30,659 --> 00:03:35,760 There are two types of diabetes type 1 86 00:03:33,300 --> 00:03:37,620 type 2. but in reality there's a whole 87 00:03:35,760 --> 00:03:39,239 lot more than that and so it actually 88 00:03:37,620 --> 00:03:41,519 helps to think of it as a Gartner's 89 00:03:39,239 --> 00:03:43,680 magic quadrant of diabetes 90 00:03:41,519 --> 00:03:46,560 so the two axes are whether you produce 91 00:03:43,680 --> 00:03:47,640 insulin and whether you are resistant to 92 00:03:46,560 --> 00:03:50,340 insulin 93 00:03:47,640 --> 00:03:52,980 so if you have type 2 which generally 94 00:03:50,340 --> 00:03:53,879 produce insulin but can't really use it 95 00:03:52,980 --> 00:03:55,739 that well 96 00:03:53,879 --> 00:03:57,420 we give the medication to reduce their 97 00:03:55,739 --> 00:03:59,819 resistance to insulin 98 00:03:57,420 --> 00:04:01,920 and you have type 1 who can use any 99 00:03:59,819 --> 00:04:03,599 insulin they have but they aren't 100 00:04:01,920 --> 00:04:05,400 producing it so we give them 101 00:04:03,599 --> 00:04:07,920 store-bought insulin 102 00:04:05,400 --> 00:04:09,959 but then as you age things change as 103 00:04:07,920 --> 00:04:11,760 people with diabetes get older they 104 00:04:09,959 --> 00:04:14,459 generally produce less insulin and 105 00:04:11,760 --> 00:04:18,000 become more resistant so we all sort of 106 00:04:14,459 --> 00:04:20,579 drift into the bottom left quadrant 107 00:04:18,000 --> 00:04:22,919 that means that some type twos will 108 00:04:20,579 --> 00:04:25,500 eventually need to take insulin and some 109 00:04:22,919 --> 00:04:27,300 type ones might need medication to lower 110 00:04:25,500 --> 00:04:29,400 their resistance to insulin 111 00:04:27,300 --> 00:04:31,800 even people who don't have diabetes tend 112 00:04:29,400 --> 00:04:33,300 to drift that way as they age so they 113 00:04:31,800 --> 00:04:34,800 may or may not eventually meet the 114 00:04:33,300 --> 00:04:36,300 criteria to be considered actually 115 00:04:34,800 --> 00:04:38,460 diabetic 116 00:04:36,300 --> 00:04:39,780 so what we want to point out at this 117 00:04:38,460 --> 00:04:41,699 stage is there are also a couple of 118 00:04:39,780 --> 00:04:44,400 in-between types which sometimes get 119 00:04:41,699 --> 00:04:45,960 called type 1.5 and they have varying 120 00:04:44,400 --> 00:04:48,540 characteristics of both 121 00:04:45,960 --> 00:04:50,220 some look more like type 1 early on and 122 00:04:48,540 --> 00:04:52,380 some look more like more like type 2 to 123 00:04:50,220 --> 00:04:54,900 start with and then change to looking 124 00:04:52,380 --> 00:04:56,639 more like type one and so we prefer to 125 00:04:54,900 --> 00:04:59,220 tend to refer to people diabetes not as 126 00:04:56,639 --> 00:05:03,320 type 1 or type 2 but as insulin 127 00:04:59,220 --> 00:05:03,320 requiring or non-insulin requiring 128 00:05:03,540 --> 00:05:07,979 so now we've got that bit out of the way 129 00:05:05,639 --> 00:05:10,740 I'd like to give you a potted history of 130 00:05:07,979 --> 00:05:13,800 my own diagnoses so I was first 131 00:05:10,740 --> 00:05:17,100 diagnosed at age 22 as a type 2 which 132 00:05:13,800 --> 00:05:19,139 surprised everyone me included and so I 133 00:05:17,100 --> 00:05:21,360 started taking a bunch of medication to 134 00:05:19,139 --> 00:05:24,180 lower my insulin resistance 135 00:05:21,360 --> 00:05:27,180 after about seven years of not really 136 00:05:24,180 --> 00:05:28,560 managing things very well I had a 137 00:05:27,180 --> 00:05:30,479 specialist suggest that perhaps I was 138 00:05:28,560 --> 00:05:32,660 one of the in-between types called Modi 139 00:05:30,479 --> 00:05:32,660 mod 140 00:05:33,479 --> 00:05:37,860 um so to get diagnosed as Modi you need 141 00:05:35,880 --> 00:05:39,539 to do a bunch of genetic testing and I 142 00:05:37,860 --> 00:05:42,060 went through all that over a period of a 143 00:05:39,539 --> 00:05:44,400 few years and eventually I had an answer 144 00:05:42,060 --> 00:05:47,400 the answer came back wait for it 145 00:05:44,400 --> 00:05:47,400 inconclusive 146 00:05:47,520 --> 00:05:52,440 so my official diagnosis then became 147 00:05:50,039 --> 00:05:54,180 atypical diabetic which isn't 148 00:05:52,440 --> 00:05:56,039 particularly useful to anyone 149 00:05:54,180 --> 00:05:58,320 but the benefit of that was that I 150 00:05:56,039 --> 00:06:01,680 managed to switch from medication to 151 00:05:58,320 --> 00:06:03,660 insulin now that might seem like a step 152 00:06:01,680 --> 00:06:06,360 backwards to go from a handful of pills 153 00:06:03,660 --> 00:06:09,360 each day to multiple daily injections 154 00:06:06,360 --> 00:06:12,000 but I can tell you it changed my life 155 00:06:09,360 --> 00:06:13,979 the plot twist to all this is that late 156 00:06:12,000 --> 00:06:18,660 last year 157 00:06:13,979 --> 00:06:21,780 um there I had a um sorry no I lost my 158 00:06:18,660 --> 00:06:24,120 place late last year I've been diabetic 159 00:06:21,780 --> 00:06:26,220 for a full 20 years at this point and 160 00:06:24,120 --> 00:06:27,960 they decided to retest for a bunch of 161 00:06:26,220 --> 00:06:30,479 antibodies that they'd never previously 162 00:06:27,960 --> 00:06:31,620 tested me for and I am now officially 163 00:06:30,479 --> 00:06:34,919 done 164 00:06:31,620 --> 00:06:37,620 type one so I've now got the full set 165 00:06:34,919 --> 00:06:40,460 basically uh insert face palm here 166 00:06:37,620 --> 00:06:40,460 basically 167 00:06:40,560 --> 00:06:45,680 oh 168 00:06:42,120 --> 00:06:45,680 let's get to the fun part the data 169 00:06:46,080 --> 00:06:50,280 so in the beginning way before both of 170 00:06:49,199 --> 00:06:52,620 our times 171 00:06:50,280 --> 00:06:54,180 there was ping on a stick that really 172 00:06:52,620 --> 00:06:55,800 just told you if your sugars were high 173 00:06:54,180 --> 00:06:58,020 enough that you might need to consider 174 00:06:55,800 --> 00:07:00,240 packing a bag to go to the hospital not 175 00:06:58,020 --> 00:07:01,680 really much more than that not much data 176 00:07:00,240 --> 00:07:03,479 there to really speak of so let's move 177 00:07:01,680 --> 00:07:06,060 on 178 00:07:03,479 --> 00:07:09,120 when I was 179 00:07:06,060 --> 00:07:11,100 I was I relied on the blood test 180 00:07:09,120 --> 00:07:14,100 machine so I won't hand it around but 181 00:07:11,100 --> 00:07:16,139 I'm sure you've all seen similar things 182 00:07:14,100 --> 00:07:19,319 and the thing that goes along with the 183 00:07:16,139 --> 00:07:21,780 blood test machine is the paper log book 184 00:07:19,319 --> 00:07:23,580 you can still walk around into just 185 00:07:21,780 --> 00:07:24,900 about any chemist and ask for one of 186 00:07:23,580 --> 00:07:29,759 these they'll give it to you for nothing 187 00:07:24,900 --> 00:07:31,919 we've just lost our systems back 188 00:07:29,759 --> 00:07:33,419 so we did we walked into my local 189 00:07:31,919 --> 00:07:35,220 chemist and we asked for a paper log 190 00:07:33,419 --> 00:07:36,419 book and they gave us one and I think we 191 00:07:35,220 --> 00:07:38,460 can probably hand that into the audience 192 00:07:36,419 --> 00:07:40,199 to pass around so people can take a 193 00:07:38,460 --> 00:07:42,300 closer look 194 00:07:40,199 --> 00:07:44,099 uh so you literally test your blood a 195 00:07:42,300 --> 00:07:47,639 few times a day it should be six to 196 00:07:44,099 --> 00:07:49,740 eight times roughly on this schedule 197 00:07:47,639 --> 00:07:51,900 uh every time you do a test you write 198 00:07:49,740 --> 00:07:55,380 the number down in your little book but 199 00:07:51,900 --> 00:07:56,759 that's just six to eight readings at 200 00:07:55,380 --> 00:07:59,479 points in time you have no information 201 00:07:56,759 --> 00:08:02,039 about the trends at any moment so 202 00:07:59,479 --> 00:08:03,479 depending on which book you get and 203 00:08:02,039 --> 00:08:04,919 unfortunately this little book that's 204 00:08:03,479 --> 00:08:06,419 going around the room doesn't have one 205 00:08:04,919 --> 00:08:08,460 but some of them actually had a little 206 00:08:06,419 --> 00:08:10,740 thing in the back where you could plot 207 00:08:08,460 --> 00:08:13,259 your graph for the day which was kind of 208 00:08:10,740 --> 00:08:15,180 neat that's kind of Handy 209 00:08:13,259 --> 00:08:17,460 um of course the whole writing it down a 210 00:08:15,180 --> 00:08:20,819 little book is prone to some error 211 00:08:17,460 --> 00:08:22,199 say for example you pull out a 12.2 a 212 00:08:20,819 --> 00:08:23,460 very high number first thing in the 213 00:08:22,199 --> 00:08:26,220 morning you're feeling a little bit 214 00:08:23,460 --> 00:08:27,840 grumpy about it call it an 8.2 or just 215 00:08:26,220 --> 00:08:29,400 forget to write it down No One's Gonna 216 00:08:27,840 --> 00:08:31,020 know right 217 00:08:29,400 --> 00:08:34,320 right 218 00:08:31,020 --> 00:08:37,020 keep that in mind we'll come back to it 219 00:08:34,320 --> 00:08:38,820 so when Lana first got her smartphone 220 00:08:37,020 --> 00:08:40,919 one of the first apps she went looking 221 00:08:38,820 --> 00:08:42,240 for was a way to store her glucose 222 00:08:40,919 --> 00:08:43,919 numbers and to throw that little 223 00:08:42,240 --> 00:08:45,660 paperwork in the bin which is why we 224 00:08:43,919 --> 00:08:47,580 don't have our original one 225 00:08:45,660 --> 00:08:50,100 the main one that she used was called 226 00:08:47,580 --> 00:08:51,540 glucosio which was open source it still 227 00:08:50,100 --> 00:08:52,740 exists although it looks a little 228 00:08:51,540 --> 00:08:54,959 abandoned 229 00:08:52,740 --> 00:08:57,120 it was nothing really more than an app 230 00:08:54,959 --> 00:08:59,700 version of the paper book with the added 231 00:08:57,120 --> 00:09:02,220 benefit of some little graphs but it 232 00:08:59,700 --> 00:09:04,800 didn't Rock a world 233 00:09:02,220 --> 00:09:07,620 then Along Came what they call a flash 234 00:09:04,800 --> 00:09:10,620 glucose monitoring system and it did 235 00:09:07,620 --> 00:09:12,120 Rock My World it also cost nearly two 236 00:09:10,620 --> 00:09:13,740 hundred dollars a month 237 00:09:12,120 --> 00:09:15,240 uh it's called the freestyle liberates 238 00:09:13,740 --> 00:09:18,300 made by pharmaceutical company called 239 00:09:15,240 --> 00:09:20,399 Abbott it's basically a sticky patch 240 00:09:18,300 --> 00:09:22,320 that you put on your body and there are 241 00:09:20,399 --> 00:09:25,920 strict rules about where on your body 242 00:09:22,320 --> 00:09:28,440 you can put it which I regularly flouted 243 00:09:25,920 --> 00:09:31,080 the device has a little filament sits 244 00:09:28,440 --> 00:09:33,720 under your skin and it would mostly but 245 00:09:31,080 --> 00:09:34,680 not always stay put there for a full 14 246 00:09:33,720 --> 00:09:36,779 days 247 00:09:34,680 --> 00:09:38,700 the filament doesn't go deep enough to 248 00:09:36,779 --> 00:09:40,459 read glucose from your blood so it 249 00:09:38,700 --> 00:09:44,100 actually reads from interstitial fluid 250 00:09:40,459 --> 00:09:46,440 the sensor itself contains not much more 251 00:09:44,100 --> 00:09:48,000 than a chip a small battery and an NFC 252 00:09:46,440 --> 00:09:50,220 transmitter 253 00:09:48,000 --> 00:09:52,320 so you hold your phone or a dedicated 254 00:09:50,220 --> 00:09:54,420 reader a receiver 255 00:09:52,320 --> 00:09:56,220 like this 256 00:09:54,420 --> 00:09:58,920 um that's available by the way at the 257 00:09:56,220 --> 00:10:01,140 low low cost of about 180 dollars 258 00:09:58,920 --> 00:10:03,120 so you can hold that up to the sensor on 259 00:10:01,140 --> 00:10:04,980 your body and it reads all the data that 260 00:10:03,120 --> 00:10:07,080 the sensor has stored and it's up to 261 00:10:04,980 --> 00:10:08,760 about eight hours so the trick there is 262 00:10:07,080 --> 00:10:10,800 don't sleep in on a Sunday morning 263 00:10:08,760 --> 00:10:13,700 because you'll lose all your data from 264 00:10:10,800 --> 00:10:13,700 late Saturday night 265 00:10:13,860 --> 00:10:19,019 have it produce an app to go with their 266 00:10:16,560 --> 00:10:22,980 sensor imagine it'll be called the 267 00:10:19,019 --> 00:10:25,680 freestyle Library link and it's fine I 268 00:10:22,980 --> 00:10:27,779 guess it has a very pretty graph 269 00:10:25,680 --> 00:10:30,060 however Lana couldn't add any extra 270 00:10:27,779 --> 00:10:32,519 information to it so if she did a blood 271 00:10:30,060 --> 00:10:34,260 test she couldn't see that and she had 272 00:10:32,519 --> 00:10:36,060 nowhere to track whether or not she 273 00:10:34,260 --> 00:10:37,740 didn't have any carbs and how many how 274 00:10:36,060 --> 00:10:39,540 much insurance he had 275 00:10:37,740 --> 00:10:42,899 but the weirdest part about this app 276 00:10:39,540 --> 00:10:44,459 that it was sometimes rewrite history at 277 00:10:42,899 --> 00:10:46,260 first I actually thought I was losing my 278 00:10:44,459 --> 00:10:48,959 mind 279 00:10:46,260 --> 00:10:51,180 um I'd scan I'd say a reading of 8.4 280 00:10:48,959 --> 00:10:53,399 then an hour later I'd look and my 281 00:10:51,180 --> 00:10:55,440 sugars hadn't gone above 7.9 282 00:10:53,399 --> 00:10:58,260 and it actually took a whole lot of time 283 00:10:55,440 --> 00:10:59,700 and weirdness and Googling and I 284 00:10:58,260 --> 00:11:02,100 actually worked out eventually that they 285 00:10:59,700 --> 00:11:03,140 have a proprietary I've lost my spot 286 00:11:02,100 --> 00:11:06,300 again 287 00:11:03,140 --> 00:11:09,000 a proprietary patented data smoothing 288 00:11:06,300 --> 00:11:11,279 algorithm that they use so when you have 289 00:11:09,000 --> 00:11:12,779 a few past readings the algorithm looks 290 00:11:11,279 --> 00:11:14,579 at those and then it Smooths out all the 291 00:11:12,779 --> 00:11:16,920 the Peaks and troughs 292 00:11:14,579 --> 00:11:19,380 um it makes your graphs look less spiky 293 00:11:16,920 --> 00:11:21,360 but it really actually just weirded me 294 00:11:19,380 --> 00:11:23,399 out it felt like the app was telling me 295 00:11:21,360 --> 00:11:26,519 that numbers I'd gotten I'd never really 296 00:11:23,399 --> 00:11:28,740 actually gotten so surely there had to 297 00:11:26,519 --> 00:11:31,560 be a way to read the data from the 298 00:11:28,740 --> 00:11:33,720 sensor and graph that directly without 299 00:11:31,560 --> 00:11:35,820 the algorithm being involved 300 00:11:33,720 --> 00:11:38,160 so from the Italian free software 301 00:11:35,820 --> 00:11:40,680 Dimension came glimp 302 00:11:38,160 --> 00:11:43,079 now glimp was a bit of a revelation it 303 00:11:40,680 --> 00:11:46,140 gave a reading and a giant red Trend 304 00:11:43,079 --> 00:11:48,180 Arrow so New View heading up and down on 305 00:11:46,140 --> 00:11:49,440 how quickly and whether or not you need 306 00:11:48,180 --> 00:11:51,420 to panic 307 00:11:49,440 --> 00:11:53,519 it also had the ability to let you type 308 00:11:51,420 --> 00:11:55,500 in your insulin data and it would show 309 00:11:53,519 --> 00:11:57,180 you how long it was active for 310 00:11:55,500 --> 00:12:00,420 Lana takes two different types of 311 00:11:57,180 --> 00:12:02,040 insulin one that's long-acting and kicks 312 00:12:00,420 --> 00:12:04,320 in an hour or so after she takes it and 313 00:12:02,040 --> 00:12:06,959 hangs around eating up glucose for 314 00:12:04,320 --> 00:12:09,540 somewhere close to 24 hours there's also 315 00:12:06,959 --> 00:12:11,940 short acting which kicks in after about 316 00:12:09,540 --> 00:12:14,279 five minutes and hangs around for a few 317 00:12:11,940 --> 00:12:15,720 hours and being able to track these 318 00:12:14,279 --> 00:12:17,519 things made a huge difference to being 319 00:12:15,720 --> 00:12:19,560 able to work out how much insulin and 320 00:12:17,519 --> 00:12:21,899 food she needed and more importantly 321 00:12:19,560 --> 00:12:24,600 when 322 00:12:21,899 --> 00:12:25,680 the problem with glimp is that it was 323 00:12:24,600 --> 00:12:28,579 just 324 00:12:25,680 --> 00:12:28,579 very Italian 325 00:12:31,019 --> 00:12:35,760 everything was in Italian 326 00:12:33,120 --> 00:12:38,459 the instructions the support presumably 327 00:12:35,760 --> 00:12:40,620 the code comments to get any kind of 328 00:12:38,459 --> 00:12:43,560 Glimpse support are usually resorted to 329 00:12:40,620 --> 00:12:48,660 just browsing Reddit it wasn't 330 00:12:43,560 --> 00:12:50,339 bad necessarily it was just janky so any 331 00:12:48,660 --> 00:12:51,959 sort of problem and I would fairly 332 00:12:50,339 --> 00:12:53,880 quickly discover that I was all on my 333 00:12:51,959 --> 00:12:55,740 own I was sort of swimming in this 334 00:12:53,880 --> 00:12:58,860 Italian diabetes world that I had no 335 00:12:55,740 --> 00:13:01,860 understanding of so for example at one 336 00:12:58,860 --> 00:13:04,740 point I started getting this random Java 337 00:13:01,860 --> 00:13:06,959 socket timeout error sometimes for hours 338 00:13:04,740 --> 00:13:09,240 on answering no readings for hours just 339 00:13:06,959 --> 00:13:11,160 a timeout error I never actually did 340 00:13:09,240 --> 00:13:12,480 work out what was causing that or how to 341 00:13:11,160 --> 00:13:14,940 fix it 342 00:13:12,480 --> 00:13:17,040 the second main thing Glenn did though 343 00:13:14,940 --> 00:13:18,240 was it introduce Lana to the world of 344 00:13:17,040 --> 00:13:19,980 night Scout 345 00:13:18,240 --> 00:13:23,279 so some of you may have gone to Dana 346 00:13:19,980 --> 00:13:25,440 Lewis's keynote at LCA a few years ago 347 00:13:23,279 --> 00:13:27,180 and so we're hoping that name isn't 348 00:13:25,440 --> 00:13:29,760 really new to you 349 00:13:27,180 --> 00:13:31,200 but if it is the basic rundown is a 350 00:13:29,760 --> 00:13:33,480 bunch of free software people got 351 00:13:31,200 --> 00:13:35,880 together and wrote a tool that allows 352 00:13:33,480 --> 00:13:38,459 you to gather glucose data from a number 353 00:13:35,880 --> 00:13:40,680 of different sources and send that on to 354 00:13:38,459 --> 00:13:42,600 a number of other different places this 355 00:13:40,680 --> 00:13:45,240 is really useful if you have an insulin 356 00:13:42,600 --> 00:13:47,339 pump so the idea is you take the data 357 00:13:45,240 --> 00:13:48,660 from your glucose monitoring system use 358 00:13:47,339 --> 00:13:50,760 night scale to have a good hard think 359 00:13:48,660 --> 00:13:53,399 about it and then send that information 360 00:13:50,760 --> 00:13:55,320 to your pump to dose you 361 00:13:53,399 --> 00:13:57,480 now if you can do all that without human 362 00:13:55,320 --> 00:13:58,440 intervention it's called a closed loop 363 00:13:57,480 --> 00:14:00,839 system 364 00:13:58,440 --> 00:14:03,540 but Lana doesn't use an insulin pump she 365 00:14:00,839 --> 00:14:06,300 has multiple daily injections 366 00:14:03,540 --> 00:14:08,160 so so to some extent she had ignored 367 00:14:06,300 --> 00:14:09,720 night Scout because it didn't seem 368 00:14:08,160 --> 00:14:11,399 particularly relevant 369 00:14:09,720 --> 00:14:13,740 but all of a sudden we had a bunch of 370 00:14:11,399 --> 00:14:15,120 engineering data and no real way to 371 00:14:13,740 --> 00:14:17,519 access it or do anything particularly 372 00:14:15,120 --> 00:14:19,620 useful with that so when she started 373 00:14:17,519 --> 00:14:21,360 using glimp one of the things she could 374 00:14:19,620 --> 00:14:23,760 do is upload it to night Scout to get 375 00:14:21,360 --> 00:14:26,100 sensor readings in a database that lived 376 00:14:23,760 --> 00:14:27,959 on cloud software that we paid in 377 00:14:26,100 --> 00:14:30,180 control for and could at least in theory 378 00:14:27,959 --> 00:14:32,519 be accessed in a bunch of new and 379 00:14:30,180 --> 00:14:33,899 interesting ways like having glucose 380 00:14:32,519 --> 00:14:36,720 data on a laptop 381 00:14:33,899 --> 00:14:40,079 or a watch 382 00:14:36,720 --> 00:14:41,940 then Abbott had a brainwave 383 00:14:40,079 --> 00:14:44,579 they had a new sensor and it was going 384 00:14:41,940 --> 00:14:46,320 to be awesome except it wasn't going to 385 00:14:44,579 --> 00:14:48,360 work with my handheld reader the one in 386 00:14:46,320 --> 00:14:50,579 the purple case that's in the room and 387 00:14:48,360 --> 00:14:52,079 it's not going to work with glimp and 388 00:14:50,579 --> 00:14:54,779 that meant that I couldn't get the data 389 00:14:52,079 --> 00:14:56,639 into night Scout anymore and by then 390 00:14:54,779 --> 00:14:58,459 I've actually addicted to having the 391 00:14:56,639 --> 00:15:00,899 data on my watch 392 00:14:58,459 --> 00:15:03,720 so we started to look for new options 393 00:15:00,899 --> 00:15:06,420 the best options seem to be Dexcom which 394 00:15:03,720 --> 00:15:08,940 came in at a cost of about 350 a month 395 00:15:06,420 --> 00:15:09,660 or almost double the cost of the Abbott 396 00:15:08,940 --> 00:15:12,240 one 397 00:15:09,660 --> 00:15:14,279 I stuck with the older Abbott sensor for 398 00:15:12,240 --> 00:15:16,019 as long as I could and as soon as they 399 00:15:14,279 --> 00:15:18,420 announced that it was death it was dead 400 00:15:16,019 --> 00:15:22,160 it was discontinued we jumped into those 401 00:15:18,420 --> 00:15:22,160 expensive Dexcom Waters 402 00:15:22,680 --> 00:15:25,680 shiny 403 00:15:25,800 --> 00:15:31,680 uh this is the Dexcom G6 it's not a 404 00:15:28,980 --> 00:15:33,000 swish Private Business Jet even if it 405 00:15:31,680 --> 00:15:35,880 feels like it costs about the same 406 00:15:33,000 --> 00:15:38,459 amount sometimes it's a complete 407 00:15:35,880 --> 00:15:41,820 continuous glucose monitoring system so 408 00:15:38,459 --> 00:15:44,279 everything I had tried before the G6 was 409 00:15:41,820 --> 00:15:46,199 really only a partial solution this was 410 00:15:44,279 --> 00:15:49,260 the real deal 411 00:15:46,199 --> 00:15:50,940 so the Dexcom has two main components a 412 00:15:49,260 --> 00:15:53,339 sensor which is the bit that goes under 413 00:15:50,940 --> 00:15:56,399 your skin and does the measuring 414 00:15:53,339 --> 00:15:58,680 and a transmitter which clips into the 415 00:15:56,399 --> 00:16:00,420 sensor on top and has a Bluetooth chip 416 00:15:58,680 --> 00:16:02,760 in it to talk to your phone 417 00:16:00,420 --> 00:16:05,220 our first question of course what's 418 00:16:02,760 --> 00:16:07,320 talking to which bit and it turns out 419 00:16:05,220 --> 00:16:09,600 nobody's allowed to know that we even 420 00:16:07,320 --> 00:16:11,699 went so far as to try and pry open a 421 00:16:09,600 --> 00:16:12,959 transmitter to have a look inside it and 422 00:16:11,699 --> 00:16:14,459 all that got was both of us having 423 00:16:12,959 --> 00:16:17,339 colored pans 424 00:16:14,459 --> 00:16:19,079 uh what we think happens is the sensor 425 00:16:17,339 --> 00:16:21,660 reads more or less every five minutes 426 00:16:19,079 --> 00:16:23,639 and sends it to the transmitter which 427 00:16:21,660 --> 00:16:24,839 stores it for some unknown period of 428 00:16:23,639 --> 00:16:26,220 time although we have a theory it's 429 00:16:24,839 --> 00:16:28,199 about three hours 430 00:16:26,220 --> 00:16:29,940 the phone then tries to handshake with 431 00:16:28,199 --> 00:16:32,339 the transmitter and authenticate with 432 00:16:29,940 --> 00:16:35,940 the code that the sensors got 433 00:16:32,339 --> 00:16:37,560 uh the transmitter auth or doesn't and 434 00:16:35,940 --> 00:16:40,019 completes their handshake and the phone 435 00:16:37,560 --> 00:16:42,120 says have you got any data for me 436 00:16:40,019 --> 00:16:44,820 if defined in the transmitter connected 437 00:16:42,120 --> 00:16:47,279 successfully five minutes ago that will 438 00:16:44,820 --> 00:16:48,839 be just one reading but if they've been 439 00:16:47,279 --> 00:16:51,060 out of contact for whatever reason that 440 00:16:48,839 --> 00:16:52,560 it'll send whatever you have so you can 441 00:16:51,060 --> 00:16:54,660 get backfilled data 442 00:16:52,560 --> 00:16:56,699 our best guess is all backfilled to 443 00:16:54,660 --> 00:16:58,880 about three hours but not much more than 444 00:16:56,699 --> 00:16:58,880 that 445 00:16:59,579 --> 00:17:04,799 of course all this is driven by phone 446 00:17:02,339 --> 00:17:06,839 and that requires an app and this is 447 00:17:04,799 --> 00:17:10,140 where we found my first problem 448 00:17:06,839 --> 00:17:12,780 Dexcom support a super limited number of 449 00:17:10,140 --> 00:17:14,880 phones and at the time I was using a 450 00:17:12,780 --> 00:17:16,260 cheap Chinese phone so absolutely no 451 00:17:14,880 --> 00:17:18,959 dice there 452 00:17:16,260 --> 00:17:20,819 in fact if you want to use the Dexcom at 453 00:17:18,959 --> 00:17:23,100 the proprietary Dexcom app on an Android 454 00:17:20,819 --> 00:17:25,919 at all you need to have an older model 455 00:17:23,100 --> 00:17:29,220 pixel or Samsung or one of exactly three 456 00:17:25,919 --> 00:17:31,679 older LG models you literally cannot 457 00:17:29,220 --> 00:17:32,520 even install the app on any other model 458 00:17:31,679 --> 00:17:35,640 phone 459 00:17:32,520 --> 00:17:39,059 so if you want to say check it out 460 00:17:35,640 --> 00:17:41,760 before you throw 300 down on this sorry 461 00:17:39,059 --> 00:17:43,140 out of luck uh luckily it's not open 462 00:17:41,760 --> 00:17:45,840 source people have worked this one out 463 00:17:43,140 --> 00:17:48,240 too and Along Came the catchly named 464 00:17:45,840 --> 00:17:50,820 extra Plus 465 00:17:48,240 --> 00:17:52,320 install version it brought with it a 466 00:17:50,820 --> 00:17:56,160 whole host of new and interesting 467 00:17:52,320 --> 00:17:58,140 problems and data so much data 468 00:17:56,160 --> 00:17:59,460 unfortunately because it needs to do 469 00:17:58,140 --> 00:18:01,440 things that Google would rather you 470 00:17:59,460 --> 00:18:03,780 didn't it's not available on the Play 471 00:18:01,440 --> 00:18:06,240 Store or even through alternative 472 00:18:03,780 --> 00:18:08,700 installation sources like F Droid so you 473 00:18:06,240 --> 00:18:11,100 need to side load it anyway we got it 474 00:18:08,700 --> 00:18:13,080 installed bought the first sensor cried 475 00:18:11,100 --> 00:18:14,460 about a bank balance and set about 476 00:18:13,080 --> 00:18:16,080 getting it going 477 00:18:14,460 --> 00:18:19,380 and the first thing you notice about X 478 00:18:16,080 --> 00:18:21,600 group is it has options lots and lots of 479 00:18:19,380 --> 00:18:24,120 options and getting them right is not 480 00:18:21,600 --> 00:18:25,799 entirely trivial we spent many hours 481 00:18:24,120 --> 00:18:28,020 just looking through the settings screen 482 00:18:25,799 --> 00:18:30,600 fiddling with things so once I actually 483 00:18:28,020 --> 00:18:33,360 managed to pocket dial a new setting 484 00:18:30,600 --> 00:18:34,740 it took me about half an hour we were 485 00:18:33,360 --> 00:18:36,660 looking through the settings we were 486 00:18:34,740 --> 00:18:38,100 Googling we were searching for stuff you 487 00:18:36,660 --> 00:18:41,100 can there's a search function in extra 488 00:18:38,100 --> 00:18:42,419 app to extract to find settings took a 489 00:18:41,100 --> 00:18:43,740 good half an hour to work out how to 490 00:18:42,419 --> 00:18:44,460 turn it off again just throw a pocket 491 00:18:43,740 --> 00:18:47,160 dial 492 00:18:44,460 --> 00:18:48,960 it also has some quirks like when you 493 00:18:47,160 --> 00:18:50,700 change setting I've been chained sensors 494 00:18:48,960 --> 00:18:52,200 it helps if you take the transmitter off 495 00:18:50,700 --> 00:18:54,720 and put it in the microwave for 10 496 00:18:52,200 --> 00:18:58,220 minutes well like for the Faraday cage 497 00:18:54,720 --> 00:18:58,220 effect not the bombarded with radiation 498 00:18:58,760 --> 00:19:03,960 that gets expensive 499 00:19:02,100 --> 00:19:06,299 so then you take it out put it in the 500 00:19:03,960 --> 00:19:08,220 new center little things like that make 501 00:19:06,299 --> 00:19:10,140 life in our house pretty exciting 502 00:19:08,220 --> 00:19:11,340 because we know how to party yeah you 503 00:19:10,140 --> 00:19:13,820 always check the microwave before you 504 00:19:11,340 --> 00:19:13,820 turn it on 505 00:19:17,820 --> 00:19:22,860 for interesting stuff so things like 506 00:19:20,820 --> 00:19:25,500 being able to snooze my alarms which my 507 00:19:22,860 --> 00:19:28,080 Dexcom app can't do the Dexcom app sorry 508 00:19:25,500 --> 00:19:30,059 I've never actually installed it it can 509 00:19:28,080 --> 00:19:32,039 send a text message to Steve with my 510 00:19:30,059 --> 00:19:33,840 location if I don't acknowledge an alarm 511 00:19:32,039 --> 00:19:35,280 within a set period of time so it'll 512 00:19:33,840 --> 00:19:37,740 basically say your girlfriend's at this 513 00:19:35,280 --> 00:19:39,960 GPS location maybe go find her 514 00:19:37,740 --> 00:19:42,120 that's also not a thing that Dexcom does 515 00:19:39,960 --> 00:19:44,520 and it also does all sorts of neat 516 00:19:42,120 --> 00:19:46,440 calculations based on insulin and carbs 517 00:19:44,520 --> 00:19:47,880 so it gives me an idea of how much I 518 00:19:46,440 --> 00:19:49,440 should take how much insulin I should 519 00:19:47,880 --> 00:19:51,900 take to counteract that massive bowl of 520 00:19:49,440 --> 00:19:53,280 pasta I ate for dinner last night or how 521 00:19:51,900 --> 00:19:56,460 many Jelly Beans I should consider 522 00:19:53,280 --> 00:19:59,160 having for a low all of which Dexcom you 523 00:19:56,460 --> 00:20:01,020 guessed it does not do then it takes all 524 00:19:59,160 --> 00:20:03,360 that juicy data and uploads it to my 525 00:20:01,020 --> 00:20:05,880 night Scout site where I can grab it see 526 00:20:03,360 --> 00:20:07,559 it on my watch on my desktop or on any 527 00:20:05,880 --> 00:20:10,340 other fancy device that I might decide 528 00:20:07,559 --> 00:20:10,340 to try out in the future 529 00:20:10,500 --> 00:20:15,000 oh Megan to get things back to the data 530 00:20:13,559 --> 00:20:16,679 I wanted to share some actual 531 00:20:15,000 --> 00:20:18,780 screenshots from the various apps that 532 00:20:16,679 --> 00:20:21,059 I've used showing some of the issues we 533 00:20:18,780 --> 00:20:23,340 have with data in different systems 534 00:20:21,059 --> 00:20:25,320 so we don't have photos of Lana's 535 00:20:23,340 --> 00:20:28,020 original paper book which probably tells 536 00:20:25,320 --> 00:20:29,520 you just how little she cared for it but 537 00:20:28,020 --> 00:20:31,799 this is the sample data and the one that 538 00:20:29,520 --> 00:20:33,000 you've you've seen uh it's not very 539 00:20:31,799 --> 00:20:35,039 interesting but it's here just for 540 00:20:33,000 --> 00:20:37,440 completeness 541 00:20:35,039 --> 00:20:40,440 this is the only screenshot I apparently 542 00:20:37,440 --> 00:20:42,960 took of glucosio despite using it for at 543 00:20:40,440 --> 00:20:45,780 least a year or two my main memory of 544 00:20:42,960 --> 00:20:47,640 the app is that it's very pink 545 00:20:45,780 --> 00:20:49,440 um we looked in the app store for it and 546 00:20:47,640 --> 00:20:50,520 it does seem to be dead so we went to 547 00:20:49,440 --> 00:20:52,500 try and find out if it's still being 548 00:20:50,520 --> 00:20:55,200 developed and it looks as though it's 549 00:20:52,500 --> 00:20:57,360 been pretty much abandoned blood test 550 00:20:55,200 --> 00:20:59,039 machines including my current one all 551 00:20:57,360 --> 00:21:00,480 her bluetooth enabled so they all 552 00:20:59,039 --> 00:21:02,700 actually come with their own proprietary 553 00:21:00,480 --> 00:21:05,039 apps the one for this one is terrible 554 00:21:02,700 --> 00:21:06,780 uh so I sort of assuming that the 555 00:21:05,039 --> 00:21:09,539 world's kind of moved on from needing an 556 00:21:06,780 --> 00:21:11,340 open source solution to to the the blood 557 00:21:09,539 --> 00:21:13,559 test machine 558 00:21:11,340 --> 00:21:15,600 so this is the proprietary habit app the 559 00:21:13,559 --> 00:21:17,520 Libra link and on a good day it would 560 00:21:15,600 --> 00:21:19,020 look like the graph on the left 561 00:21:17,520 --> 00:21:21,059 but then there are the other days where 562 00:21:19,020 --> 00:21:24,000 there'd be gaps and weird persistent 563 00:21:21,059 --> 00:21:26,940 lows that weren't actually lows like the 564 00:21:24,000 --> 00:21:28,500 one on the right those we usually caused 565 00:21:26,940 --> 00:21:29,940 by the sets are coming loose on the 566 00:21:28,500 --> 00:21:32,400 under skin and the little filament 567 00:21:29,940 --> 00:21:34,799 pulling out in some cases this could be 568 00:21:32,400 --> 00:21:36,600 a little painful we experimented a lot 569 00:21:34,799 --> 00:21:40,460 with with different tapes to hold Those 570 00:21:36,600 --> 00:21:40,460 sensors on for the full 14 days 571 00:21:42,480 --> 00:21:47,820 app which of course I can't install so 572 00:21:45,299 --> 00:21:50,640 we asked my mum who uses it to send me 573 00:21:47,820 --> 00:21:52,559 some screenshots unfortunately the app 574 00:21:50,640 --> 00:21:54,900 also has a security setting that means 575 00:21:52,559 --> 00:21:56,700 you can't screenshot it so this is a 576 00:21:54,900 --> 00:21:58,559 cropped photograph of my mum holding her 577 00:21:56,700 --> 00:21:59,940 phone 578 00:21:58,559 --> 00:22:03,480 um you can see it has this weird 579 00:21:59,940 --> 00:22:05,280 teardrop here so this this Arrow part 580 00:22:03,480 --> 00:22:08,220 points up or down depending on whether 581 00:22:05,280 --> 00:22:10,200 you're heading up or down and so it sort 582 00:22:08,220 --> 00:22:10,919 of points up or down moves during the 583 00:22:10,200 --> 00:22:12,299 day 584 00:22:10,919 --> 00:22:13,380 now in this shot it's pointing to the 585 00:22:12,299 --> 00:22:15,600 right which means that sugars are 586 00:22:13,380 --> 00:22:18,240 holding steady so hi I might add 587 00:22:15,600 --> 00:22:20,340 bitstick 588 00:22:18,240 --> 00:22:22,380 and this is extra 589 00:22:20,340 --> 00:22:24,720 which on a good day would look like this 590 00:22:22,380 --> 00:22:26,820 on the left you can see the blue dots 591 00:22:24,720 --> 00:22:29,159 are readings the green dots being 592 00:22:26,820 --> 00:22:31,320 insulin on board and the purple dots 593 00:22:29,159 --> 00:22:33,600 being a prediction based on readings 594 00:22:31,320 --> 00:22:35,280 carbs and insulin 595 00:22:33,600 --> 00:22:37,140 and at the top of the screen is usually 596 00:22:35,280 --> 00:22:39,840 a calculation saying what it thinks Lana 597 00:22:37,140 --> 00:22:42,120 should do in terms of carbs or insulin 598 00:22:39,840 --> 00:22:44,100 sometimes exciting stuff happens like 599 00:22:42,120 --> 00:22:46,080 this on the right where it has a purple 600 00:22:44,100 --> 00:22:48,480 Direction line going up 601 00:22:46,080 --> 00:22:50,520 saying as well as the red prediction 602 00:22:48,480 --> 00:22:52,679 line going down in this case it's 603 00:22:50,520 --> 00:22:54,600 telling Anna to take insulin but with a 604 00:22:52,679 --> 00:22:57,740 warning symbol which is extra version of 605 00:22:54,600 --> 00:22:57,740 the shrug Emoji like 606 00:22:59,640 --> 00:23:04,140 so then you have days when you put a new 607 00:23:01,620 --> 00:23:06,600 sensor on and your body kind of freaks 608 00:23:04,140 --> 00:23:08,940 out a little you are essentially 609 00:23:06,600 --> 00:23:10,559 creating a little wound in your skin and 610 00:23:08,940 --> 00:23:13,080 it actually takes a day or two for your 611 00:23:10,559 --> 00:23:14,760 body to get used to it being there all 612 00:23:13,080 --> 00:23:18,120 of which means your readings go a little 613 00:23:14,760 --> 00:23:20,100 Haywire for about 24 hours so this one 614 00:23:18,120 --> 00:23:22,460 is from the Libre on the left and you 615 00:23:20,100 --> 00:23:26,220 can see it's just a bit Wild 616 00:23:22,460 --> 00:23:27,720 the one on the right is X strip and I 617 00:23:26,220 --> 00:23:29,100 mean admittedly I'm eating food and 618 00:23:27,720 --> 00:23:30,600 taking insulin and running around and 619 00:23:29,100 --> 00:23:33,299 doing all that stuff as well but that 620 00:23:30,600 --> 00:23:35,100 that data is noisy 621 00:23:33,299 --> 00:23:36,659 so when things get really exciting in 622 00:23:35,100 --> 00:23:38,340 excerpt you have readings like the one 623 00:23:36,659 --> 00:23:41,340 on the left with the double arrow down 624 00:23:38,340 --> 00:23:42,659 which is basically eat all the sugar and 625 00:23:41,340 --> 00:23:45,059 do it fast 626 00:23:42,659 --> 00:23:46,860 or this one on the right where you can 627 00:23:45,059 --> 00:23:48,659 see Lana has had quite the day eating 628 00:23:46,860 --> 00:23:51,000 all the carbs and consequently injecting 629 00:23:48,659 --> 00:23:53,820 all of the insulin 630 00:23:51,000 --> 00:23:55,260 so cake is delicious so let's run 631 00:23:53,820 --> 00:23:57,900 through how things are currently working 632 00:23:55,260 --> 00:23:59,880 we've broken this stuff down in timing 633 00:23:57,900 --> 00:24:03,620 because it gives you an idea of just how 634 00:23:59,880 --> 00:24:03,620 much work is involved in managing this 635 00:24:03,900 --> 00:24:08,940 so starting with ratings readings happen 636 00:24:06,360 --> 00:24:11,580 more or less every five minutes with my 637 00:24:08,940 --> 00:24:14,220 old phone it was closer to eight with my 638 00:24:11,580 --> 00:24:15,360 new phone it's closer to four so the 639 00:24:14,220 --> 00:24:18,780 phone actually does make a difference 640 00:24:15,360 --> 00:24:21,480 and how frequent your readings are 641 00:24:18,780 --> 00:24:23,640 um either way it's around 250 to 300 642 00:24:21,480 --> 00:24:25,559 readings a day which is a far cry from 643 00:24:23,640 --> 00:24:28,260 the six to eight blood tests I was doing 644 00:24:25,559 --> 00:24:31,380 when I was first diagnosed 645 00:24:28,260 --> 00:24:33,960 diet scalp Paul's extra every few 646 00:24:31,380 --> 00:24:37,020 minutes and update the database with any 647 00:24:33,960 --> 00:24:39,000 new data so let us watch desktop widget 648 00:24:37,020 --> 00:24:42,059 browser plugins all then pull from night 649 00:24:39,000 --> 00:24:43,380 Scout on the road schedules and we 650 00:24:42,059 --> 00:24:44,700 haven't read the code for that but most 651 00:24:43,380 --> 00:24:46,740 of them are probably around five minutes 652 00:24:44,700 --> 00:24:48,840 as well 653 00:24:46,740 --> 00:24:49,740 insulin injections happen multiple times 654 00:24:48,840 --> 00:24:52,020 a day 655 00:24:49,740 --> 00:24:54,000 I take long acting morning and night at 656 00:24:52,020 --> 00:24:55,500 roughly the same time every day although 657 00:24:54,000 --> 00:24:57,840 I can screw it up by an hour or two or 658 00:24:55,500 --> 00:25:02,100 forget completely like I did yesterday 659 00:24:57,840 --> 00:25:03,780 and nothing terrible really happens 660 00:25:02,100 --> 00:25:06,240 um 661 00:25:03,780 --> 00:25:07,860 so short acting is more ad hoc if I'm 662 00:25:06,240 --> 00:25:09,480 about to have a giant bowl of rice I'll 663 00:25:07,860 --> 00:25:11,120 have an injection as I sit down to eat 664 00:25:09,480 --> 00:25:13,620 and then maybe another one later on 665 00:25:11,120 --> 00:25:15,539 pasta and potatoes work differently as 666 00:25:13,620 --> 00:25:17,280 does bread and simple sugars like 667 00:25:15,539 --> 00:25:19,260 lollies cake or chocolate 668 00:25:17,280 --> 00:25:20,580 it's also important to take into account 669 00:25:19,260 --> 00:25:22,860 exercise 670 00:25:20,580 --> 00:25:25,140 some exercise sends me high while I'm 671 00:25:22,860 --> 00:25:28,200 doing it some sends me low 672 00:25:25,140 --> 00:25:29,700 all exercises send me low afterwards so 673 00:25:28,200 --> 00:25:31,559 I usually need to make sure I'm within 674 00:25:29,700 --> 00:25:32,820 arm's length of some carbs after a 675 00:25:31,559 --> 00:25:34,860 workout 676 00:25:32,820 --> 00:25:36,539 it's also important to note that other 677 00:25:34,860 --> 00:25:38,100 things can affect whether or not I need 678 00:25:36,539 --> 00:25:40,380 insulin as well so it's not just food 679 00:25:38,100 --> 00:25:42,260 and exercise things like stress your 680 00:25:40,380 --> 00:25:44,580 mood the weather weight loss and gain 681 00:25:42,260 --> 00:25:46,440 incidental exercise like gardening or 682 00:25:44,580 --> 00:25:48,720 walking around the shops and how much 683 00:25:46,440 --> 00:25:51,120 overall exercising I say how many times 684 00:25:48,720 --> 00:25:53,760 I've been to the gym this week not just 685 00:25:51,120 --> 00:25:56,100 whether or not I'm there now all those 686 00:25:53,760 --> 00:25:58,200 things have a can make a difference 687 00:25:56,100 --> 00:26:00,299 and so then there's a dawn effect 688 00:25:58,200 --> 00:26:03,299 one thing that human bodies do to wake 689 00:26:00,299 --> 00:26:05,580 us up is our liver says oh hi and gives 690 00:26:03,299 --> 00:26:07,980 you a little hit of glucose so if you 691 00:26:05,580 --> 00:26:09,539 have a working pancreas you process that 692 00:26:07,980 --> 00:26:11,100 and you start feeling hungry and then 693 00:26:09,539 --> 00:26:12,840 you go and get breakfast 694 00:26:11,100 --> 00:26:14,400 people with diabetes instead tend to 695 00:26:12,840 --> 00:26:16,380 wake up with high blood sugars 696 00:26:14,400 --> 00:26:18,779 regardless of what they've been eating 697 00:26:16,380 --> 00:26:20,700 or what readings they had overnight and 698 00:26:18,779 --> 00:26:22,620 that means either living with the high 699 00:26:20,700 --> 00:26:24,059 numbers and skipping breakfast or having 700 00:26:22,620 --> 00:26:26,520 some insulin so that you can have a 701 00:26:24,059 --> 00:26:28,940 piece of toast 702 00:26:26,520 --> 00:26:31,500 we didn't test this so brace yourself 703 00:26:28,940 --> 00:26:34,140 alarms happen more or less every day 704 00:26:31,500 --> 00:26:36,740 during the day you have a low alarm 705 00:26:34,140 --> 00:26:39,859 which sounds like this 706 00:26:36,740 --> 00:26:39,859 [Music] 707 00:26:40,080 --> 00:26:43,200 oh good 708 00:26:41,460 --> 00:26:45,679 um and uh should we do that again yeah 709 00:26:43,200 --> 00:26:45,679 okay 710 00:26:49,760 --> 00:26:54,740 and a high alarm which sounds like this 711 00:26:57,600 --> 00:27:01,200 if you've been in a talk after Lana's 712 00:26:59,820 --> 00:27:02,520 had lunch you might have heard one of 713 00:27:01,200 --> 00:27:04,559 those already there's been at least a 714 00:27:02,520 --> 00:27:07,559 couple listening 715 00:27:04,559 --> 00:27:10,620 so the lows go off under 3.8 and the 716 00:27:07,559 --> 00:27:13,620 high alarms that go off over 8.0 717 00:27:10,620 --> 00:27:17,100 at night the alarm is quieter and calmer 718 00:27:13,620 --> 00:27:18,720 but a little bit more insistent and less 719 00:27:17,100 --> 00:27:19,820 hilarious when it's two o'clock in the 720 00:27:18,720 --> 00:27:22,440 morning 721 00:27:19,820 --> 00:27:23,760 but the range is also a little wider 722 00:27:22,440 --> 00:27:25,919 because we don't want to get woken up 723 00:27:23,760 --> 00:27:28,679 unless we really need to do something 724 00:27:25,919 --> 00:27:30,659 about it I'll say Lana because I don't 725 00:27:28,679 --> 00:27:32,880 usually wake up 726 00:27:30,659 --> 00:27:35,039 so this is another bonus of x-trip over 727 00:27:32,880 --> 00:27:37,260 the Dexcom app because we have these 728 00:27:35,039 --> 00:27:41,039 complicated alarm settings with custom 729 00:27:37,260 --> 00:27:43,500 sounds and custom schedules 730 00:27:41,039 --> 00:27:46,260 so on a blood test they do still happen 731 00:27:43,500 --> 00:27:48,659 uh they happen to regularly now so not 732 00:27:46,260 --> 00:27:50,640 six to eight times a day I will usually 733 00:27:48,659 --> 00:27:52,860 do a few during the two hour startup 734 00:27:50,640 --> 00:27:54,840 window with a new sensor but I'll also 735 00:27:52,860 --> 00:27:56,400 do one whenever I look at the number my 736 00:27:54,840 --> 00:27:58,140 sensor is giving me and think it just 737 00:27:56,400 --> 00:27:59,400 doesn't seem quite right I mean 738 00:27:58,140 --> 00:28:01,140 sometimes I'm the one who's wrong and 739 00:27:59,400 --> 00:28:04,140 it's fine but other times it is a little 740 00:28:01,140 --> 00:28:06,179 bit inaccurate so extra allows me to 741 00:28:04,140 --> 00:28:08,340 calibrate based on blood test readings 742 00:28:06,179 --> 00:28:10,559 so because my blood test machine is 743 00:28:08,340 --> 00:28:12,900 Bluetooth I do a test it sends it 744 00:28:10,559 --> 00:28:14,520 directly into extra extra then says do 745 00:28:12,900 --> 00:28:16,860 you want me to calibrate I can say yes 746 00:28:14,520 --> 00:28:18,720 or no or it may fail to send it into 747 00:28:16,860 --> 00:28:20,100 extra sometimes sometimes because my 748 00:28:18,720 --> 00:28:21,360 machine's a bit terrible it fails to 749 00:28:20,100 --> 00:28:24,240 send index drip enough to do it manually 750 00:28:21,360 --> 00:28:26,220 which is complain worthy 751 00:28:24,240 --> 00:28:28,200 um because interstitial fluid lags 752 00:28:26,220 --> 00:28:29,940 behind blood glucose measurements X 753 00:28:28,200 --> 00:28:31,740 strip is actually smart enough to know 754 00:28:29,940 --> 00:28:33,779 that the blood test reading needs to be 755 00:28:31,740 --> 00:28:35,279 calibrated about 15 minutes into the 756 00:28:33,779 --> 00:28:38,340 future as well which I think is really 757 00:28:35,279 --> 00:28:40,200 neat in essence I will always trust a 758 00:28:38,340 --> 00:28:41,820 blood test reading over a sensor reading 759 00:28:40,200 --> 00:28:44,940 if I'm unsure 760 00:28:41,820 --> 00:28:47,100 census last 10 days they take about 15 761 00:28:44,940 --> 00:28:48,900 minutes to change over because they've 762 00:28:47,100 --> 00:28:52,020 got to go in the microwave and then they 763 00:28:48,900 --> 00:28:54,779 need to warm up for two hours because I 764 00:28:52,020 --> 00:28:56,520 don't know I guess Dexcom has reasons so 765 00:28:54,779 --> 00:28:59,059 this means the time that you change the 766 00:28:56,520 --> 00:29:01,679 census creeps later and later in the day 767 00:28:59,059 --> 00:29:03,240 uh and at some point whether they go 768 00:29:01,679 --> 00:29:04,320 overnight and we'll just switch back in 769 00:29:03,240 --> 00:29:07,140 the mornings 770 00:29:04,320 --> 00:29:08,760 and travel travel makes this even more 771 00:29:07,140 --> 00:29:11,659 hilarious 772 00:29:08,760 --> 00:29:11,659 I'm happy to hand this 773 00:29:12,240 --> 00:29:14,840 record or break 774 00:29:18,299 --> 00:29:21,960 that basically that whole plastic unit 775 00:29:20,220 --> 00:29:25,580 including the packaging gets thrown away 776 00:29:21,960 --> 00:29:25,580 and that happens every 10 days 777 00:29:26,399 --> 00:29:31,799 so I went to transmitters transmitters 778 00:29:28,740 --> 00:29:33,840 last 90 days and so far it's actually 779 00:29:31,799 --> 00:29:36,539 aligned really nicely to every ninth 780 00:29:33,840 --> 00:29:38,700 sensor we have no doubt that we will 781 00:29:36,539 --> 00:29:40,320 screw this up at some point and things 782 00:29:38,700 --> 00:29:42,720 are going to get even more exciting 783 00:29:40,320 --> 00:29:44,700 because transmitters are not designed to 784 00:29:42,720 --> 00:29:47,640 be taken out of the sensors while the 785 00:29:44,700 --> 00:29:50,640 sensor is attached to your body it is 786 00:29:47,640 --> 00:29:52,679 possible however we've done it 787 00:29:50,640 --> 00:29:54,299 uh transmitters take a little longer to 788 00:29:52,679 --> 00:29:56,100 change over they take up to about half 789 00:29:54,299 --> 00:29:57,419 an hour or so and that's because we need 790 00:29:56,100 --> 00:29:59,580 to make sure it's operating correctly 791 00:29:57,419 --> 00:30:01,200 before we kick off the sensor so we have 792 00:29:59,580 --> 00:30:03,179 to sit there and wait for it wait for it 793 00:30:01,200 --> 00:30:04,320 to report its battery voltages through X 794 00:30:03,179 --> 00:30:05,700 strip once we've seen the battery 795 00:30:04,320 --> 00:30:08,279 voltages we can then kick off the 796 00:30:05,700 --> 00:30:10,520 transmitter and we we start again start 797 00:30:08,279 --> 00:30:13,320 the next 10 days 798 00:30:10,520 --> 00:30:16,740 and and the thing on the longer schedule 799 00:30:13,320 --> 00:30:18,480 is probably the phone although Lano has 800 00:30:16,740 --> 00:30:21,539 brought exactly one new phone since 801 00:30:18,480 --> 00:30:23,399 using extra detectable neck strip and 802 00:30:21,539 --> 00:30:24,840 despite both our nerves about changing 803 00:30:23,399 --> 00:30:26,940 over between phones it actually went 804 00:30:24,840 --> 00:30:30,020 really smoothly 805 00:30:26,940 --> 00:30:30,020 who expected that 806 00:30:31,380 --> 00:30:36,380 I also bought a new watch recently 807 00:30:34,200 --> 00:30:38,880 and that was a whole lot less seamless 808 00:30:36,380 --> 00:30:40,860 despite the fact that I chose this watch 809 00:30:38,880 --> 00:30:43,620 specifically because it should have 810 00:30:40,860 --> 00:30:46,140 worked better with my setup it turns out 811 00:30:43,620 --> 00:30:47,760 not to and now I still regularly switch 812 00:30:46,140 --> 00:30:49,380 back to my old watch for exercise 813 00:30:47,760 --> 00:30:50,940 tracking because it actually gives me 814 00:30:49,380 --> 00:30:53,820 better glucose data while I'm working 815 00:30:50,940 --> 00:30:56,059 out um this one looks way prettier alone 816 00:30:53,820 --> 00:30:58,260 so you win some you lose some I guess 817 00:30:56,059 --> 00:31:02,120 now to the part we know you've all been 818 00:30:58,260 --> 00:31:02,120 waiting for the actual data 819 00:31:04,380 --> 00:31:11,159 so my original night Scout setup was a 820 00:31:07,640 --> 00:31:13,799 mongodb database with a Heroku VM and 821 00:31:11,159 --> 00:31:15,299 hosting recently Heroku ditched their 822 00:31:13,799 --> 00:31:17,640 free tier hosting and threatened to 823 00:31:15,299 --> 00:31:19,440 delete my data so we decided to redo 824 00:31:17,640 --> 00:31:21,480 everything with Google Google Cloud 825 00:31:19,440 --> 00:31:23,460 platform which means that it remains 826 00:31:21,480 --> 00:31:25,440 free and everything is moved into one 827 00:31:23,460 --> 00:31:27,659 location which is a bit more convenient 828 00:31:25,440 --> 00:31:29,760 so when we shut down the old mongodb 829 00:31:27,659 --> 00:31:31,980 database we exported a bunch of my data 830 00:31:29,760 --> 00:31:34,620 mostly because I work for a company that 831 00:31:31,980 --> 00:31:36,179 does time series databases and having a 832 00:31:34,620 --> 00:31:38,880 bunch of Time series data lying around 833 00:31:36,179 --> 00:31:40,620 could come in handy someday 834 00:31:38,880 --> 00:31:42,960 um so when we got accepted to do this 835 00:31:40,620 --> 00:31:44,700 talk we thought now was probably the 836 00:31:42,960 --> 00:31:47,279 time to upload some of this historical 837 00:31:44,700 --> 00:31:49,020 data into a time scale database run a 838 00:31:47,279 --> 00:31:51,779 few interesting queries and maybe make 839 00:31:49,020 --> 00:31:53,220 some nice graphs which sounds easy we 840 00:31:51,779 --> 00:31:54,899 also thought it would be super cool if 841 00:31:53,220 --> 00:31:56,760 we could pull down some Wellness data 842 00:31:54,899 --> 00:31:59,279 from my Garmin Smartwatch and my smart 843 00:31:56,760 --> 00:32:01,140 my smart bathroom scales and see if we 844 00:31:59,279 --> 00:32:03,360 could use the power of postgres queries 845 00:32:01,140 --> 00:32:04,980 to get some interesting data and the 846 00:32:03,360 --> 00:32:06,919 answer to all that was 847 00:32:04,980 --> 00:32:10,620 sort of 848 00:32:06,919 --> 00:32:12,659 we started with a whole bunch of CSP 849 00:32:10,620 --> 00:32:14,820 files with different columns different 850 00:32:12,659 --> 00:32:17,100 data sources different Drive stamps you 851 00:32:14,820 --> 00:32:18,899 name it then we had the exciting stuff 852 00:32:17,100 --> 00:32:21,059 like trying to concatenate that data 853 00:32:18,899 --> 00:32:23,760 from x-drip with the data from glymp 854 00:32:21,059 --> 00:32:25,260 because the cold names are different 855 00:32:23,760 --> 00:32:27,240 and then the government had an extra 856 00:32:25,260 --> 00:32:29,460 Rose into the table for weight 857 00:32:27,240 --> 00:32:31,740 where the date was on a row all of its 858 00:32:29,460 --> 00:32:34,260 own with null fields and then the time 859 00:32:31,740 --> 00:32:35,840 with the with the actual data was on the 860 00:32:34,260 --> 00:32:38,340 row underneath it 861 00:32:35,840 --> 00:32:40,440 uh so then you had to manually move it 862 00:32:38,340 --> 00:32:43,080 around it and CSV files are 863 00:32:40,440 --> 00:32:45,000 okay as a data source 864 00:32:43,080 --> 00:32:47,340 you can throw them into a spreadsheet 865 00:32:45,000 --> 00:32:48,720 with a tiny bit of effort and that might 866 00:32:47,340 --> 00:32:51,120 actually allow to graph some numbers 867 00:32:48,720 --> 00:32:53,580 that are actually present or make some 868 00:32:51,120 --> 00:32:55,080 things based on those numbers 869 00:32:53,580 --> 00:32:56,820 but if you want to query things group 870 00:32:55,080 --> 00:32:58,559 them do something else you need 871 00:32:56,820 --> 00:33:02,039 something else like a Time series 872 00:32:58,559 --> 00:33:04,260 database converting from CSV to SQL is 873 00:33:02,039 --> 00:33:05,580 tricky at the best of times since first 874 00:33:04,260 --> 00:33:08,460 you need to create a table and then 875 00:33:05,580 --> 00:33:10,440 adjust that data into the table 876 00:33:08,460 --> 00:33:12,480 an SQL databases tend to be pretty 877 00:33:10,440 --> 00:33:15,600 particular about things like data 878 00:33:12,480 --> 00:33:18,360 formats and where it sees me just quotes 879 00:33:15,600 --> 00:33:21,720 it it says that's fine isn't it 880 00:33:18,360 --> 00:33:23,100 um so you can either edit these so you 881 00:33:21,720 --> 00:33:24,120 can either transform the data as you 882 00:33:23,100 --> 00:33:26,460 ingest it 883 00:33:24,120 --> 00:33:27,600 or you can edit the SQL statements in 884 00:33:26,460 --> 00:33:29,820 the file 885 00:33:27,600 --> 00:33:33,360 which can get exciting 886 00:33:29,820 --> 00:33:35,399 so we used sqlite so because it can 887 00:33:33,360 --> 00:33:37,980 directly read CSV files to create a 888 00:33:35,399 --> 00:33:39,779 table for you and then you can dump that 889 00:33:37,980 --> 00:33:42,000 to print out the resulting SQL which you 890 00:33:39,779 --> 00:33:42,600 can see at the bottom of the slide 891 00:33:42,000 --> 00:33:44,159 um 892 00:33:42,600 --> 00:33:47,220 but 893 00:33:44,159 --> 00:33:49,919 it's all the column types of text 894 00:33:47,220 --> 00:33:51,899 which means then you can't query things 895 00:33:49,919 --> 00:33:53,340 properly you've got to cast it or you've 896 00:33:51,899 --> 00:33:55,679 got to alter the table to change the 897 00:33:53,340 --> 00:33:58,260 column type most notably the date and 898 00:33:55,679 --> 00:33:59,820 time Fields must be a times empty Zed 899 00:33:58,260 --> 00:34:01,679 field for time scale to even convert a 900 00:33:59,820 --> 00:34:03,480 table to a hyper table 901 00:34:01,679 --> 00:34:06,120 all in all this took us about three 902 00:34:03,480 --> 00:34:08,520 weekends more than a few work nights and 903 00:34:06,120 --> 00:34:10,980 a whole lot of wine and a whole lot of 904 00:34:08,520 --> 00:34:12,119 waning and it's a bit of a one that 905 00:34:10,980 --> 00:34:14,359 we're still talking about each other 906 00:34:12,119 --> 00:34:14,359 really 907 00:34:14,820 --> 00:34:18,960 we did successfully put together a 908 00:34:16,980 --> 00:34:23,060 single working query which we then 909 00:34:18,960 --> 00:34:23,060 squirted into grafana and got a graph 910 00:34:25,500 --> 00:34:30,480 so this graph shows my change in BMI 911 00:34:28,560 --> 00:34:31,919 over the course of four months or so 912 00:34:30,480 --> 00:34:34,260 while I was actively trying to lose some 913 00:34:31,919 --> 00:34:37,139 weight and it plots it against my total 914 00:34:34,260 --> 00:34:39,300 weekly insulin usage essentially the 915 00:34:37,139 --> 00:34:41,460 more insulin you take the more weight 916 00:34:39,300 --> 00:34:42,839 you gain so if you have insulin 917 00:34:41,460 --> 00:34:44,399 molecules running around in your body 918 00:34:42,839 --> 00:34:46,679 that aren't binding themselves to 919 00:34:44,399 --> 00:34:48,899 glucose you will have a nice low blood 920 00:34:46,679 --> 00:34:50,700 glucose number to show your doctor that 921 00:34:48,899 --> 00:34:53,580 the insulin is just busy storing itself 922 00:34:50,700 --> 00:34:56,399 as fat in case you need it later on now 923 00:34:53,580 --> 00:34:58,160 remember I talked earlier about lying in 924 00:34:56,399 --> 00:35:00,900 your little paper book 925 00:34:58,160 --> 00:35:02,820 people with diabetes are really prone to 926 00:35:00,900 --> 00:35:05,099 eating disorders um for hopefully 927 00:35:02,820 --> 00:35:06,599 obvious reasons and there's a fairly 928 00:35:05,099 --> 00:35:09,180 well-known phenomenon that's sometimes 929 00:35:06,599 --> 00:35:11,520 called diabolemium This is where people 930 00:35:09,180 --> 00:35:13,440 will allow their sugars to run high and 931 00:35:11,520 --> 00:35:15,720 not take insulin so that they lose 932 00:35:13,440 --> 00:35:17,700 weight and if you're lying in your 933 00:35:15,720 --> 00:35:20,220 little book you can get away with that 934 00:35:17,700 --> 00:35:22,859 for a little while you will come unstuck 935 00:35:20,220 --> 00:35:25,740 though when you get certain other longer 936 00:35:22,859 --> 00:35:27,599 range blood tests done with your GP so 937 00:35:25,740 --> 00:35:29,160 it's certainly not a good strategy and I 938 00:35:27,599 --> 00:35:31,440 don't recommend it if you're considering 939 00:35:29,160 --> 00:35:33,540 it so back to the graph 940 00:35:31,440 --> 00:35:35,280 you can see where I was restricting my 941 00:35:33,540 --> 00:35:38,520 carb intake 942 00:35:35,280 --> 00:35:40,380 so it's restricting my carb intake I 943 00:35:38,520 --> 00:35:42,359 lost my spot again in order to restrict 944 00:35:40,380 --> 00:35:44,220 my insulin intake so the insulin line is 945 00:35:42,359 --> 00:35:46,500 dropping 946 00:35:44,220 --> 00:35:49,020 and that allowed me to lose weight so my 947 00:35:46,500 --> 00:35:51,000 BMI is also dropping there 948 00:35:49,020 --> 00:35:53,160 and my BMI actually dropped quite 949 00:35:51,000 --> 00:35:55,260 rapidly and it works surprisingly well 950 00:35:53,160 --> 00:35:57,119 then I decided I was happy with my 951 00:35:55,260 --> 00:35:59,220 weight where I was for the moment I 952 00:35:57,119 --> 00:36:01,079 started adding back carbs which meant I 953 00:35:59,220 --> 00:36:03,300 was having more insulin and my weight 954 00:36:01,079 --> 00:36:05,460 loss started to level off so the insulin 955 00:36:03,300 --> 00:36:06,540 comes back up and the weight loss levels 956 00:36:05,460 --> 00:36:08,820 out so it's actually a really 957 00:36:06,540 --> 00:36:10,260 interesting graph to to ponder I think 958 00:36:08,820 --> 00:36:12,540 anyway 959 00:36:10,260 --> 00:36:14,400 so to finish up I think the main point 960 00:36:12,540 --> 00:36:16,200 we want to make in this talk is wouldn't 961 00:36:14,400 --> 00:36:18,660 it be nice if we could somehow access 962 00:36:16,200 --> 00:36:20,820 all of this data that we collect about 963 00:36:18,660 --> 00:36:23,160 ourselves between two or three different 964 00:36:20,820 --> 00:36:24,599 devices some of which even have the same 965 00:36:23,160 --> 00:36:27,060 brand names on them 966 00:36:24,599 --> 00:36:29,280 they could uh they collect sleep data 967 00:36:27,060 --> 00:36:32,940 heart rate data nutrition data weight 968 00:36:29,280 --> 00:36:35,579 data glucose data insulin data and sorts 969 00:36:32,940 --> 00:36:37,500 of other things mostly they're stored in 970 00:36:35,579 --> 00:36:38,520 a database that we can sort of export to 971 00:36:37,500 --> 00:36:40,800 CSV 972 00:36:38,520 --> 00:36:42,599 uh but to maybe even try and get a small 973 00:36:40,800 --> 00:36:44,640 amount of historical data like that 974 00:36:42,599 --> 00:36:46,020 simple graph took literal hours of 975 00:36:44,640 --> 00:36:47,820 effort 976 00:36:46,020 --> 00:36:50,160 we're smart people 977 00:36:47,820 --> 00:36:52,619 we work on this kind of thing all the 978 00:36:50,160 --> 00:36:54,500 time and we actually have some idea what 979 00:36:52,619 --> 00:36:56,880 we're talking about on a technical level 980 00:36:54,500 --> 00:36:59,460 imagine if people who had just been 981 00:36:56,880 --> 00:37:01,980 diagnosed had accessed information like 982 00:36:59,460 --> 00:37:04,320 I do that lets them know which brand of 983 00:37:01,980 --> 00:37:05,700 rice is best for them to eat or how to 984 00:37:04,320 --> 00:37:07,380 manage a walk in the park without 985 00:37:05,700 --> 00:37:09,359 killing over halfway home because you 986 00:37:07,380 --> 00:37:11,760 forgot to bring jelly beans with you 987 00:37:09,359 --> 00:37:13,560 if information wants to be free someone 988 00:37:11,760 --> 00:37:15,720 should let the data know because right 989 00:37:13,560 --> 00:37:18,000 now that's locked up in jail possibly 990 00:37:15,720 --> 00:37:20,720 permanently 991 00:37:18,000 --> 00:37:20,720 thank you 992 00:37:26,820 --> 00:37:31,800 sponsor sorry if you could answer it 993 00:37:29,760 --> 00:37:34,160 I'll take this round yeah sure so 994 00:37:31,800 --> 00:37:34,160 questions 995 00:37:39,359 --> 00:37:43,619 what do you know what's the reason for 996 00:37:41,940 --> 00:37:46,200 the uh 997 00:37:43,619 --> 00:37:47,820 transmitter to need to be replaced so I 998 00:37:46,200 --> 00:37:50,220 can kind of get why the sensor might be 999 00:37:47,820 --> 00:37:52,500 battery life battery life so it's not 1000 00:37:50,220 --> 00:37:54,599 rechargeable it is not rechargeable no 1001 00:37:52,500 --> 00:37:57,300 it lasts for 90 days and then that's it 1002 00:37:54,599 --> 00:38:00,079 it can't pull it apart and we tried to 1003 00:37:57,300 --> 00:38:02,280 pull it apart glued it's glued to glued 1004 00:38:00,079 --> 00:38:04,920 people have tried to extend it because 1005 00:38:02,280 --> 00:38:07,440 they are quite expensive but 1006 00:38:04,920 --> 00:38:09,480 web kind of privilege so we just after 1007 00:38:07,440 --> 00:38:12,000 90 days we throw it in the bin and we 1008 00:38:09,480 --> 00:38:15,440 take a new one and sigh about more 1009 00:38:12,000 --> 00:38:15,440 single-use plastic effectively 1010 00:38:16,160 --> 00:38:20,820 is is there any um 1011 00:38:19,260 --> 00:38:23,700 have 1012 00:38:20,820 --> 00:38:26,040 you ever seen any sign off or are aware 1013 00:38:23,700 --> 00:38:29,099 of anyone looking at making Open Source 1014 00:38:26,040 --> 00:38:32,240 Hardware for this no no just because 1015 00:38:29,099 --> 00:38:32,240 it's a medical device 1016 00:38:35,040 --> 00:38:39,599 there's strict legislation around 1017 00:38:36,720 --> 00:38:42,560 medical devices which is 1018 00:38:39,599 --> 00:38:42,560 horrifying 1019 00:38:47,760 --> 00:38:51,900 I noticed the extra Villa face is quite 1020 00:38:50,220 --> 00:38:53,339 busy compared to the commercial software 1021 00:38:51,900 --> 00:38:55,200 interface I was just curious if you 1022 00:38:53,339 --> 00:38:58,079 think there'd be potential for a 1023 00:38:55,200 --> 00:38:59,520 simplified free software interface or if 1024 00:38:58,079 --> 00:39:01,560 you need all of that data in order to 1025 00:38:59,520 --> 00:39:02,460 really tell what the sensor is doing and 1026 00:39:01,560 --> 00:39:04,820 tell the difference between the 1027 00:39:02,460 --> 00:39:07,560 different modes more data is more data 1028 00:39:04,820 --> 00:39:08,700 but yeah I mean yes some more simplified 1029 00:39:07,560 --> 00:39:11,099 didn't face 1030 00:39:08,700 --> 00:39:14,220 could be nice but to be honest x-trip 1031 00:39:11,099 --> 00:39:16,380 supports way more than just the G6 1032 00:39:14,220 --> 00:39:18,599 um there I I can't remember I've thought 1033 00:39:16,380 --> 00:39:20,700 of my head but it's at least uh five or 1034 00:39:18,599 --> 00:39:22,260 six different different actual 1035 00:39:20,700 --> 00:39:23,940 transmitters pumps and things as well 1036 00:39:22,260 --> 00:39:26,040 right one thing I will say though and 1037 00:39:23,940 --> 00:39:27,420 something I've heard from other people 1038 00:39:26,040 --> 00:39:30,599 in diabetes Community who are not 1039 00:39:27,420 --> 00:39:32,460 technical is when you do first get your 1040 00:39:30,599 --> 00:39:33,599 first CGM or your first monitor like 1041 00:39:32,460 --> 00:39:36,660 this the data is actually super 1042 00:39:33,599 --> 00:39:39,119 overwhelming so I love it but that's at 1043 00:39:36,660 --> 00:39:41,400 least partly because I'm a data nerd 1044 00:39:39,119 --> 00:39:43,200 um sometimes especially I think if 1045 00:39:41,400 --> 00:39:46,800 you're less technically Savvy having 1046 00:39:43,200 --> 00:39:49,079 that that data coming at you is quite 1047 00:39:46,800 --> 00:39:50,820 overwhelming I think it's Pro there's 1048 00:39:49,079 --> 00:39:52,700 probably an argument for you could have 1049 00:39:50,820 --> 00:39:55,619 a simplified interface that you could 1050 00:39:52,700 --> 00:39:58,320 unsimplify perhaps over time I love 1051 00:39:55,619 --> 00:40:00,000 having all the information both of us 1052 00:39:58,320 --> 00:40:01,680 constantly are checking on the the 1053 00:40:00,000 --> 00:40:03,300 status of my sensor and making sure that 1054 00:40:01,680 --> 00:40:04,980 my transmitter is working correctly and 1055 00:40:03,300 --> 00:40:06,780 all that kind of stuff I personally I 1056 00:40:04,980 --> 00:40:10,520 love it but I can absolutely see the 1057 00:40:06,780 --> 00:40:10,520 argument for a less complicated 1058 00:40:10,880 --> 00:40:16,820 and she's like that's nice but nope not 1059 00:40:14,160 --> 00:40:16,820 not doing that 1060 00:40:18,960 --> 00:40:21,800 heading to the back 1061 00:40:26,339 --> 00:40:31,560 hi more comment than um question 1062 00:40:29,940 --> 00:40:33,119 um being another in still independent 1063 00:40:31,560 --> 00:40:34,859 diabetic in the room 1064 00:40:33,119 --> 00:40:37,020 um with the pump the transmitter is 1065 00:40:34,859 --> 00:40:39,960 actually rechargeable 1066 00:40:37,020 --> 00:40:41,400 but these since the actual sensor gets 1067 00:40:39,960 --> 00:40:43,920 has to get replaced every seven days 1068 00:40:41,400 --> 00:40:46,880 yeah but you've got the constant graph 1069 00:40:43,920 --> 00:40:50,280 but I have the same problem as you the 1070 00:40:46,880 --> 00:40:52,260 Medtronic software does not work on my 1071 00:40:50,280 --> 00:40:54,359 phone it I think it is literally only 1072 00:40:52,260 --> 00:40:58,380 the Samsung and maybe a couple of other 1073 00:40:54,359 --> 00:41:01,140 Android phones that 1074 00:40:58,380 --> 00:41:03,540 um actually let you use it yeah foreign 1075 00:41:01,140 --> 00:41:06,359 support is a massive thing and I wish 1076 00:41:03,540 --> 00:41:09,180 that Dexcom and other manufacturers 1077 00:41:06,359 --> 00:41:11,099 actually took it more seriously but I 1078 00:41:09,180 --> 00:41:12,180 think it's time and data time time and 1079 00:41:11,099 --> 00:41:14,880 money really 1080 00:41:12,180 --> 00:41:17,280 it used to be really hard because 1081 00:41:14,880 --> 00:41:20,460 um I'd have to upload the software 1082 00:41:17,280 --> 00:41:23,460 Through The Carling thing and at one 1083 00:41:20,460 --> 00:41:26,400 point it would only work on Windows for 1084 00:41:23,460 --> 00:41:28,440 OS a Microsoft OS 1085 00:41:26,400 --> 00:41:31,280 a kid complaining and going when are you 1086 00:41:28,440 --> 00:41:31,280 getting Linux version 1087 00:41:34,920 --> 00:41:38,000 cool any more questions 1088 00:41:38,160 --> 00:41:42,740 no if we could have a final round of 1089 00:41:40,320 --> 00:41:42,740 applause 1090 00:41:42,810 --> 00:41:46,060 [Applause]