Pitfalls of Insulin Pump Clocks

Abstract

Objective:

The objective was to raise awareness about the importance of ensuring that insulin pumps internal clocks are set up correctly at all times. This is a very important safety issue because all commercially available insulin pumps are not GPS-enabled (though this is controversial), nor equipped with automatically adjusting internal clocks. Special attention is paid to how basal and bolus dose errors can be introduced by daylight savings time changes, travel across time zones, and am-pm clock errors. Correct setting of insulin pump internal clock is crucial for appropriate insulin delivery.

Methods:

A comprehensive literature review is provided, as are illustrative cases.

Results:

Incorrect setting can potentially result in incorrect insulin delivery, with potential harmful consequences, if too much or too little insulin is delivered. Daylight saving time changes may not significantly affect basal insulin delivery, given the triviality of the time difference. However, bolus insulin doses can be dramatically affected. Such problems may occur when pump wearers have large variations in their insulin to carb ratio, especially if they forget to change their pump clock in the spring. More worrisome than daylight saving time change is the am-pm clock setting. If this setting is set up incorrectly, both basal rates and bolus doses will be affected.

Conclusions:

Appropriate insulin delivery through insulin pumps requires correct correlation between dose settings and internal clock time settings. Because insulin pumps are not GPS-enabled or automatically time-adjusting, extra caution should be practiced by patients to ensure correct time settings at all times. Clinicians and diabetes educators should verify the date/time of insulin pumps during patients’ visits, and should remind their patients to always verify these settings.

Daylight saving time (DST) change occurs twice a year (spring and fall), as people switch their clocks and watches from winter’s time to summer’s time and vice versa. Some clocks adjust automatically, and some need manual adjustment. During this change, caution should be taken in various life situations, especially in the health care sector.

In this article, we discuss some precautions related to time changes in regard to insulin pump clocks, and how glitches in insulin pump clocks may affect insulin delivery, resulting in potential errors. DST change is just one example of such glitches; other issues include glitches related to am-pm modes, traveling across time zones, and other related glitches in insulin pumps’ internal clocks. While these potential errors have not been reported to cause serious concerns in the diabetes literature, it is worthwhile to be aware of these technical glitches. It is not certain, though, whether the lack of relevant literature is due to underreporting or a reality that such errors are extremely rare. However, in our diabetes practice, we have encountered few cases that we believe would be worth sharing. We encountered this clinical concern recently, and our work was summarized in a blog published by the American Association of Diabetes Educators (AADE), ¹ and we thought of reiterating this precautionary piece in this article, expanding the discussion, to raise awareness of patients and clinicians about these technical glitches.

DST has been developed as a means “of making better use of the daylight in the evenings by setting the clocks forward one hour during the longer days of summer, and back again in the fall” as defined by the Time and Date website. ² Around the world, DST is not adopted in all countries. ² In the United States, most states practice DST change. The state of Indiana only recently adopted DST (in 2005), and the states of Arizona and Hawaii do not practice DST change. The aforementioned website details the timing of the application of DST around the world, ² by country. We have noted that most countries that adopt DST apply the summer time in March or early April and winter time in late October or early November.

Although it may be considered a trivial issue by some, and it may be conceived as a simple task, but to manually change clocks during DST changes is not always an easy task, especially nowadays when numerous electronic gadgets and devices have automatic (or GPS-connected) internal clocks. As is human nature, people may not invariably pay attention to details. When it comes to insulin pumps, adjusting the pumps’ internal clocks may be overlooked by some patients when changing batteries or during DST changes. Insulin pumps are like all other electronic gadgets that have internal clocks.

As an analogy to other devices, one can imagine the hassles encountered in resetting clocks around the house (eg, alarm clocks, appliances clocks) after a power outage, or resetting the car clock when the car battery dies after inadvertently leaving the headlights on. Some clocks are difficult to reset (eg, you cannot go backward in hours or minutes, or you have to use a clip or pin to change numbers). Also, choosing the am-pm function can be annoying or tricky in some clocks. We have observed these difficulties anecdotally, in discussions with colleagues and friends. Unfortunately, insulin pump technology is lagging behind in regard to the internal pump clock programming. As will be discussed later, most insulin pumps nowadays do not automatically adjust for DST changes, and therefore it remains the responsibility of the patient to make this seasonal adjustment. The correct time/date setting is crucial for correct function of Insulin pumps (or continuous subcutaneous insulin infusion, CSII).

Insulin pumps are effective tools for insulin delivery in patients with diabetes, ³ but correct insulin delivery is dependent on correct function of the internal pump clock. The settings of insulin delivery (ie, basal and bolus settings) are programmed to occur by the hour (or part of an hour). The delivery of basal insulin occurs as a certain amount delivered per hour; in fact the pump delivers this amount in tiny fractions constantly, for example every 1-3 minutes, to maintain a preset basal rate, for example, 0.8 units per hour.

While different CSII pumps work differently, most pumps available on the market are programmed for basal rates by units per hour. Similarly, bolus infusions (for carb coverage and correction for high glucose) are set up to be delivered at a starting time. Insulin to carb ratios (I:C) can be set differently at different times of day in a pump and this can introduce significant changes in bolus dose size when a pump’s clock is improperly set. It is important to note that if a patient utilizes only a single rather than multiple basal rates and bolus doses, there will be no issues or errors in insulin doses delivered. Clock-related problems discussed in this review may result only in the case of the use of multiple basal or bolus settings. For accurate insulin delivery to occur, the pump clock should be set up correctly to align with the actual times intended for insulin doses to be delivered (both basal and bolus).

Practically speaking, while basal rates may not be significantly affected by DST changes (in view of a mere 1-hour difference), bolus insulin delivery may be affected more significantly. Currently available pumps have programs that calculate premeal doses and correction doses based on carb ratios, glucose readings and glucose targets. These are called by different names, for example, wizards or bolus calculators, and so on. Problems related to DST changes may occur when pump wearers have large variations in their I:C ratios, especially if they forget to change their pump clock in the spring. Pumpers who have a stronger (lower) I:C ratio at breakfast and a weaker (higher) one set at 11:00 am for lunch during the summer time will get a larger lunch bolus dose if they give a lunch bolus prior to noon after the spring time change has occurred.

For example, if a patient uses a 1:10 carb ratio for breakfast and a 1:15 ratio for lunch (because of risk of afternoon hypoglycemia due to light lunch and higher physical activity), then taking the 1:10 carb-ratio dose at lunch, by mistake, could result in afternoon hypoglycemia. It is prudent to point out here that some patients may not follow the aforementioned meal schedule, which is typical for Americans, according to our observations. For example, some European societies do not eat meals on the same schedules as Americans do. Therefore, it is prudent to time the wizard settings according to the mealtimes, if using such automated wizards. It is obvious that this discussion applies only to situations where patients use these automated programs (wizards), allowing the pump to calculate the bolus doses based on the aforementioned input data (in relation to carb ratios and correction factors). These doses are governed by correct clock timings. However, if patients do not use these wizards, but rather calculate their doses manually, then such precautions about clock settings will not apply, regardless of the time of meals.

The aforementioned hypothetical case scenario could occur as follows: Traditionally, but variably, carb ratios are set up in insulin pumps at midnight, 11:00 am, and 5:00 pm to coincide with the times of breakfast, lunch, and dinner, respectively. These times are set up because most people generally have lunch around 11:30 am to 12:00 pm and dinner around 5:30 to 6:00 pm, while breakfast time is more flexible for most people. During the spring DST change, the lunch time will advance, and so if the carb ratio time on the pump clock is not manually adjusted, the morning (1:10) carb ratio will be delivered at lunch, delivering a 33% bigger insulin dose (unless the patient picks up the glitch when the pump asks the patient to verify the final dose calculation). Other hypothetical bolus dose miscalculations could occur as a result of not adjusting the pump clock for the DST change.

am-pm time setting is another issue of paramount importance for CSII. This latter setting is certainly more problematic than the DST seasonal change: This setting can affect both basal and bolus insulin delivery. If a patient has significantly different basal rates for early morning (eg, much lower rates for midnight to 6:00 am vs noon to 6:00 pm) and the patient happens to have times mixed up in the pump clock, he or she will actually get the higher rates at night/early morning, and this can potentially result in nocturnal hypoglycemia.

Similarly, premeal doses could be affected by incorrect am-pm settings. For example if a patient uses a 1:10 carb ratio for dinner and a 1:15 ratio for lunch (bigger vs smaller insulin doses, respectively), and the patient inputs 60 grams of carbs, the pump will calculate 6 units for the 1:10 carb ratio and 4 units for the 1:15 carb ratio. If the am and pm settings are not correct, the patient will receive incorrect boluses at lunch and dinner (risk of postmeal hypo- or hyperglycemia, respectively).

To illustrate how this hypothetical am-pm setting error may occur, resulting in undesirable outcomes, an analogy comes to mind, in a different domain (aviation traffic): In December 2013, a complete shutdown of air traffic occurred in the United Kingdom, paralyzing airports around the country. This was major news in Europe and elsewhere because this air traffic shutdown affected air traffic around the world, since UK airports are transit ports for numerous international airlines. The cause was a glitch in the communication computer system that had to do with the system’s clock. The system, located at England’s main traffic control center in Swanwick, Hampshire, failed to switch from night- to daytime. In the health care domain we strive to avoid similar IT glitches.

While the above discussion about clock settings may be perceived as being merely hypothetical in the case of health care settings, we would like to share the following case of a patient we encountered recently. As summarized in the AADE’s blog, ¹ “we encountered a patient with type 1 diabetes (T1DM) who manifested a glitch related to internal pump’s clock setting. This 18 year old international college student with T1DM had been using an insulin pump, which was started in her native country (where no DST changes occur). When she was seen in the clinic in late summer (2012), her pump’s clock was 1 hour late, because she did not realize the need to manually change the clock in the preceding spring. No significant harm resulted from this incorrect setting.” To raise awareness about this glitch, we presented this case as an abstract/poster at the AADE 2012 annual meeting in Indianapolis, Indiana. This was well received by the meeting attendants.

Interestingly, there is not much published literature about these issues related to insulin pump clock settings, which we believe is due to underestimation of the potential clock glitches or underreporting of adverse events. We recently performed a thorough literature search and found only brief discussions of insulin pump time settings.^4-6 Issues related to travel across time zones ⁴ and the quest for synchronization between different diabetes-related IT gadgets ⁶ were emphasized. It is prudent to emphasize the former issue in regard to travel across time zones ⁴ alluded to above. This will be further discussed later.

We also have recently reviewed the printed user’s manuals accompanying commercially available insulin pumps^7-11 focusing on instructions related to pump internal clocks. We found that none of the insulin pumps’ internal clocks adjust automatically for DST seasonal changes. Likewise, there are no contemporary pump versions that have built-in GPS technologies, which would thus ensure self-adjustment of time/date settings at all times. However, as will be further discussed, the issue of implementing GPS technology into insulin pumps is controversial for various reasons, with expected skepticism by some, and therefore will remain a debatable issue.

All manuals for commercially available insulin pumps (in the United States) instruct patients about the importance of ensuring correct clock settings at all times, in variable details.^7-11 One pump manual instructs users in regard to DST ⁷ that “if they spring forward after 11:00 pm but before midnight, they will also have to manually change the date. If they wait until after midnight, the pump date will have changed automatically to the appropriate date.” Likewise, the manual recommends that the wearers set the clock back before midnight on Saturday or after 1:00 am on Sunday to keep the pump set to the correct date/time. It is interesting that the manual instructs pump wearers that they will register an additional hour in the daily totals history, because the day has been altered to contain 25 hours. If the clock is changed between midnight and 1:00 am the date will have to be manually changed. The history will then contain a duplicate date and contain up to one hour of insulin delivered. ⁷

Another pump’s manual ⁸ warns that incorrect setting of the time and date can lead to incorrect basal delivery and data memory. The warning states, “After inserting a new battery, always check that the time and date. . . . Incorrect programming of the time and/or date may cause incorrect insulin delivery. If you, your doctor or healthcare team store and analyze your therapy data electronically and the time and date of the devices used are not set identically, the resulting data may not be meaningful.” ⁸ Another recommendation by another pump manual was to emphasize time/date verification when deactivating the pump or when insulin delivery is suspended. ⁹

Of note, we came across these 2 glitches:

1. Leap year setting. A memo was issued by a pump manufacturer ¹² regarding a “glitch” related to its pump software not recognizing the 2012 leap date (February 29). The warning stated that “an incorrect time may result in basal insulin delivery or a recommended bolus amount that is incorrect for the actual time of day,” and that “this could result in unexpected high or low blood sugars.” While this leap year glitch has not been reported to cause significant errors in insulin pumps, we ourselves encountered a relevant case where a patient wearing that pump brand was seen in the following month of March, and we found her pump’s clock to be off by one day (March began on February 29), but no serious adverse events occurred due to this glitch!?

2. Clock expiration (End of 2015): In early 2013, the US Food and Drug Administration (FDA) circulated a recall of a batch of insulin pumps ¹³ that would no longer function after December 31, 2015. The FDA’s recall document ¹³ stated that “due to an internal calendar date limit (December 31, 2015) these pumps will cease to operate on January 1, 2016, and will result in a call service alarm that is displayed on the pump display.” These pumps will continue to function normally until December 31, 2015, but we would like to remind all patients and clinicians that all users of this batch of pumps should replace their pumps prior to that date (prior to December 31, 2015) ¹³ .

Finally, we would like to allude to an important issue that can similarly result in incorrect insulin delivery. This is in regard to setting the starting and ending times for the basal rates, whether the setting is changed by patients or by their providers. While different brands have different setting procedures for basal rates, it is prudent to pay attention to infusion time periods. We recently encountered a case of a patient with T1DM on an insulin pump, coming for his first visit to our diabetes center. While one of our providers was making changes to basal rates as appropriate, it was noted that the period of 10:00 pm to midnight was set at 0 units/hour, meaning that the patient was receiving no insulin for 2 hours.

While this was just a 2-hour period of inadvertent suspense of insulin delivery, such human error could potentially result in similar errors but for longer periods. In T1DM, this can be harmful, if prolonged enough (eg, no insulin for several hours during the night), which would be compounded by dawn hyperglycemia, and could potentially result in diabetic ketoacidosis. Upon further questioning, the patient recalled having changed his basal rates recently. Acknowledging that he may have accidently left that period at “0” units/hour, no serious harm resulted.

Based on the aforementioned discussion, diabetes education guidelines should include a section about pump clock settings and potential glitches. We propose that diabetes organizations take a look at these issues and evaluate the issue further. Until then we suggest the following solutions:

1. It is prudent that patients set up the pumps’ time and date correctly at all times; upon initial use, upon changing batteries, and whenever a pump acts up, raising questions about pump software jamming, and so on.

2. It is important to point out that clinicians may not notice that the am and pm settings are switched when glucose readings are reviewed. This can lead to egregious clinical errors such as raising the night basal accidentally for daytime hyperglycemia due to the clock error. Therefore, diabetes specialists, diabetes nurses, and diabetes educators should always verify the am-pm settings on insulin pumps (and glucose meters for that matter) when reviewing/downloading glucose logs.

3. To obviate human errors in regard to the am-pm settings, one suggestion is to use the 24-hour format, that is, military timing (eg, 18:00 for 6:00 pm). However, we acknowledge that some people may not be familiar with this 24-hour format. Furthermore, we also encountered a case where a patient was using military time on his pump, but the pump software installed in our clinic’s computer could not download his settings and data, due to incorrect synchronization of time settings.

4. At each patient’s encounter, diabetes specialists, diabetes nurses, and diabetes educators should verify the pump’s time/date, as well as the starting/ending dates of insulin administration. We admit that we ourselves may sometimes forget to verify the time/date settings of patients’ pumps and glucose meters. Clinicians should familiarize themselves with all available insulin pump brands, since different brands use different setting methods.

5. It is prudent to elaborate on the issue of travel across time zones, alluded to briefly, earlier. ⁴ Kaufman and Westfall explained that the human body may take some time to adjust to new time zones. ⁴ Travel across time zones poses special problems for people with diabetes, including encountering irregular schedules, different food choices, additional physical activity, and possibly additional stress related to travel mishaps. Also, timing of meals will change while traveling. Therefore, basal and mealtime insulin dosing must be adjusted to accommodate all of the aforementioned changes, stemming from crossing multiple time zones. It is interesting to note that (to our knowledge) such travel-induced changes are encountered while traveling east–west and vice versa, but not vertically (ie, north–south and vice versa). Furthermore, the human body needs a few days (up to 4 days) to adjust to new day–night (circadian) changes, which may trigger changes in hormonal function that will ultimately influence glucose homeostasis.

6. It is noted that pump manufacturers strive to educate patients and clinicians through instruction manuals, memos, and other forms of communication, about all safety issues related to insulin pumps, including clock settings. An example of this continued effort by manufacturers, beyond instruction manuals,^{7 -11} is this manufacturer’s warning ¹⁴ to pump users that was posted on the manufacturer’s website in 2012 in regard to DST, stating, “Your device will not automatically update the time for daylight saving time.” However, to prevent such technical pitfalls completely, more steps are required of pump manufacturers in future pumps. We recommend that pump manufactures consider the following solutions:

7. For the future, manufacturers are encouraged to equip insulin pumps with GPS connectivity: This will guarantee that pumps’ clocks automatically adjust at all times, including DST adjustment and automatic adjustment during travel across time zones. Another advantage of GPS is to help with identifying patients’ location in cases of emergency (a rarely needed situation, but not totally imaginary). However, there are potential logistical and technological obstacles. For example, a concern may be patients’ data security. In fact, we anticipate strong opposition to this GPS-connectivity proposal by many skeptics. Eavesdropping/spying and personal privacy are major (and genuine) concerns that are expressed by the public whenever the issue of IT and wireless/satellite communications is discussed. Therefore, we call for further discussion of this issue among health care and diabetes experts. Furthermore, it is conceivable that societies around the world may vary in their acceptance/rejection of the privacy vulnerability imposed by GPS connectivity. Therefore, we recommend that this issue be discussed country-wise.

8. We acknowledge that the above 2 suggestions may be challenged by the unique physiological adjustment of travel across time zones, as discussed by Kaufman and Westfall. ⁴ As discussed above (point 4), this concept implies that patients may want to gradually change their pump settings according to multiple cofactors related to travel that may be a hindrance to automatic clock adjustment. However, Crowe and Klonoff advocated automated time adjustment, including implementing GPS technologies, stating, “And all of you who are developing promising technologies that are time-dependent, make sure you are watching the clock!” ⁶

To conclude, we would like to emphasize that until all pumps should be equipped with automatic time adjustment (or GPS if feasible) and that it is prudent that diabetes clinicians, nurses, and educators educate their patients about adjustments for seasonal time and date changes (DST, leap year, year 2016) as well as to pay close attention to current clock settings, for example, pm versus am. Furthermore, just as important, diabetes clinicians, nurses, and educators should educate their patients to constantly pay attention to current date and time settings of their pumps and to constantly verify that the pump clocks are set up correctly at all times.