Future Artificial Pancreas Technology for Type 1 Diabetes: What Do Users Want?

Abstract

Background:

The primary focus of artificial pancreas (AP) research has been on technical achievements, such as time in range for glucose levels or prevention of hypoglycemia. Few studies have attempted to ascertain the expectations of users of AP technology.

Subjects and Methods:

Persons with type 1 diabetes and parents of children with type 1 diabetes were invited to take part in an online survey concerning future use and expectations of AP technology. The survey was advertised via Twitter, Facebook, and DiabetesMine, plus advocacy groups and charities including INPUT, Diabetes UK, and the Diabetes Research and Wellness Foundation. Quantitative responses were categorized on a 5-point Likert scale. Free text responses were analyzed using content analysis.

Results:

Two hundred sixty-six surveys were completed over a 1-month period. Two hundred forty participants indicated they were highly likely to use a fully automated 24-h AP. Approximately half of the respondents indicated they would be likely to use a device that only functioned overnight. Size, visibility, and lack of effectiveness were the top reasons for not wanting an AP. Despite perceived potential downsides, participants expressed a strong need for a device that will help minimize the burden of disease, help facilitate improved psychosocial functioning, and improve quality of life.

Conclusions:

Background

Artificial pancreas (AP) research has progressed significantly over recent years, raising expectations of availability of the technology within the next 3–5 years.¹ It is particularly important during the development of AP systems to understand and manage these expectations and to understand the psychosocial implications of this technology in order to aid user engagement.² The primary aim of current AP research is to create a fully automated, 24-h closed-loop system with minimal or no input by the user. For people with diabetes, this will require confidence in the system's ability to perform to achieve their desired goals but at a potential “cost” of loss of personal control. Furthermore, any AP must fit with an individual's lifestyle and not add unnecessary burden of disease to an already burdensome experience.

The first AP system that becomes available will probably coordinate an insulin pump, a continuous glucose monitor (CGM), and a “controller” (the brains of the AP system), which might end up as an application (app) on a smartphone with time. The controller coordinates inputs from the CGM to determine how much insulin to give to achieve euglycemia. It must also account for intended carbohydrate intake, previous insulin doses given (to calculate insulin-on-board), and possibly other inputs to direct insulin dosing from the insulin pump. Some systems will have the controller built into the insulin pump to help avoid communication problems between the devices, with the phone app as a means to interface with the controller. A fully automated 24 h-a-day AP system means that the user would need to wear an insulin pump and CGM, as well as carry the controller. Insulin pump cannulae would need to be replaced every 2–3 days, and CGM sensors would need to be replaced every 5–7 days. Owing to limitations of the speed of onset of even the most rapid-acting subcutaneous insulins currently available, most AP systems in testing today continue to require meal announcement.

The aim of the current project was to gain a greater understanding of the views of potential future AP users.

Participants and Methods

People with type 1 diabetes and parents of children with type 1 diabetes were invited to take part in an online survey. Questions were developed among the multidisciplinary team and people with diabetes, prior to piloting the draft survey with potential participants. Minor revisions were made in line with feedback prior to “going live” with the survey.

The electronic survey provided a brief explanation of the basic concept of the AP and then explored views and preferences of adults with type 1 diabetes and parents of children with type 1 diabetes. It was hosted on the University of Southampton isurvey site over a 1-month period following ethics approval from Santa Barbara Cottage Hospital Ethics Board (dated August 12, 2014). Advertising occurred via the social media sites Twitter, Facebook, and DiabetesMine, as well as patient advocacy groups including INPUT, Diabetes UK, and Diabetes Research and Wellness Foundation charities and investigators' healthcare professional networks.

Quantitative responses were analyzed using SPSS statistical software version 21 (SPSS, Inc., Chicago, IL). Free text responses were analyzed using content analysis.

Results

The number of participants was 266, of whom 168 were female. Participant mean age was 34.2 years (range, 3–74 years), with a mean diabetes duration of 19.5 years (range, 7 weeks to 68 years).

Insulin pump therapy was used by 204 participants, with 111 participants using CGM devices. Of these, 79 used a CGM all of the time. For those who did not use a CGM all the time, the main reason for use was to gain “better control of my diabetes” (n=22) and for exercise (n=14). Other reasons included pregnancy (n=3), overnight (n=2), and to fine-tune glycemic control (n=2). The majority of participants (n=177) wanted an AP system that tries to keep them to a specific blood glucose level (for example, 90 mg/dL [5 mmol/L]), whereas others (n=86) preferred one that kept them to a range (for example, 70–180 mg/day [4–10 mmol/L]).

When asked, “overall, if it were available, how likely is it that you would want to use a fully automated 24 hours a day artificial pancreas to manage your diabetes?,” 240 participants said extremely or very likely. For an AP that worked only overnight, only 105 participants said they would be likely to use it, with 97 saying they would not or would be very unlikely to use it, with a further 57 undecided.

Responses were similar for use of a system that required the user to “tell it” that he or she was planning to eat or exercise by pressing a button but not how many carbohydrates or duration of exercise. Two hundred forty-two participants were extremely or highly likely to use such a system. If the user was required to provide the additional information of number of carbohydrates consumed or duration/intensity of exercise, the number of those extremely or very likely to use the system dropped to 165.

Regarding thinking about situations where an AP device would be particularly helpful, the top responses were people who have hypoglycemia unawareness (n=135), are pregnant (n=46), and for everyone (n=29). Other responses included exercise (n=25), children/young people (n=20), people with specific challenges such as disability (n=16), and those newly diagnosed (n=11).

Specific comments provided greater clarification and highlighted the challenges faced by participants. Quotes include “Hypos in the night, illness, stress, hormonal changes…. Life really and living without all the other stuff you have to carry with you” and “I struggle to control my diabetes on a day to day basis so would find it helpful in everyday situations.” Further comments included “I think it would be very beneficial for small children with unpredictable eating patterns” and “Pretty much anything!”

Table 1 shows all responses to the survey question “What would a successful artificial pancreas look like for you?” Table 2 details all responses to the survey question “What would stop you from using an artificial pancreas if it were available?” Responses were similar irrespective of therapy or CGM use for both questions. Table 3 reports survey responses to agreement on potential advantages of an AP system. Similarly, there were no differences between people who use a CGM or the broader participant population in response to these answers. Table 4 illustrates the separated data for CGM users only.

Table 1.

Survey Responses to “What Would a Successful Artificial Pancreas Look Like for You?”

Response	n	%
Small	85	31.9
Discreet/appearance/no canula	55	20.7
Effective (BG in range/dual hormone/like a pancreas)	42	15.8
Minimal patient intervention/low maintenance	38	14.3
All in one device	33	12.4
Like an insulin pump	20	7.5
I don't care, I just want it to work	14	5.3
Accurate/reliable	11	4.1
Only one canula/sensor	11	4.1
Flexible to fit lifestyle/versatile/freedom to live my life	11	4.1
Allows me to have a normal life/better QoL/happiness/not have to think about diabetes all the time/avoid db exhaustion	9	3.4
Like a smartphone or MP3	7	2.6
Large visual display/audio alarm	3	1.1
Is implantable	4	1.5
Something that I can override	2	0.8
Anything that allowed me to sleep	2	0.8
Something that integrates within my other diabetes devices	2	0.8
Easy to use	2	0.8
Cheap/affordable	2	0.8
Not painful	2	0.8
Has a longer durability of sensor and transmitter	1	0.4
Robust	1	0.4

“Other” responses included “ANYTHING that allows me to sleep through the night” and “I could have a normal life.”

BG, blood glucose; db, diabetes; QoL, quality of life.

Table 2.

Survey Responses to “What Would Stop You from Using an Artificial Pancreas If It Were Available?”

Response	n	%
Cost	93	35.0
Size/appearance/constant attachment/potential scarring	63	23.7
Lack of reliability/accuracy	31	11.7
Nothing	28	10.5
Impracticality/inflexibility	18	6.8
Functionality	16	6.0
Too much personal input required	15	5.6
Ineffective/problematic algorithm	8	3.0
Reliance on a machine/loss of personal control	7	2.6
Safety	7	2.6
No perceived additional benefit	6	2.3
Pain	5	1.9
Allergic reaction to adhesive/cannulas	5	1.9
Prototype/early stage development (wait for others to try first)	3	1.1
Too steep a learning curve/daunting	2	0.8
Lack of clinical support	3	1.1
A cure	1	0.4
AK47 trained at my head	1	0.4
If it only worked for 2 years and had to stop, emotional impact	1	0.4

“Other” responses included “If accuracy could not be guaranteed,” “If I had to tell it what I was doing. I might as well just manage on my own then,” “Possibly cost. Other than that I can't see anything stopping me,” “If it were very bulky, if it was more painful to place than existing pump and if it had side effects or if it had to be repositioned every day,” “If the CGM [continuous glucose monitor] can't read blood sugar correctly,” “I wouldn't want to lose complete awareness of diabetes in case the AP [artificial pancrease] broke and I needed to go back to a pump,” “If I had several gadgets to keep track off or had to wear multiple devices/CGM plus 2 sites I wouldn't be interested,” “If it restricted my normal day to day living,” “If it was cumbersome, too many wires/cannulas overtly complicated,” “Large allergic reaction with adhesive tape or infusion sets/inserts, lack of flexibility in coping with spontaneous changes in exercise plan or food intake or changes in BG [blood glucose] due to other factors,” and “Controlling it for maintaining normal glycemia takes too long or is too difficult to operate.”

Table 3.

Survey Responses to Agreement on Potential Advantages of an Artificial Pancreas System

			n
Question	Total	Average	1	2	3	4	5
Improved overall blood glucose control measured by HbA1c levels?	264	4.6	2	6	21	46	189
Fewer low glucose levels (i.e., <70 mg/dL or 4 mmol/L)?	262	4.5	0	7	26	63	166
Fewer high glucose levels (i.e., >180 mg/dL or 10 mmol/L)?	263	4.6	1	6	12	58	186
Less day-to-day swings in your blood glucose levels?	261	4.6	1	2	17	60	181
Less worry about the long-term effects of your diabetes?	263	4.5	2	5	31	56	169
Time off from thinking about your diabetes so much?	262	4.1	8	19	43	48	144
Less worry for your family members when they think about your diabetes?	263	3.6	25	32	63	59	84
Less interference from diabetes in everyday life?	264	4.2	6	14	42	66	136
A better night's sleep?	261	3.8	13	31	55	56	106
Starting the day on a “good” blood sugar number?	263	4.1	2	9	57	86	109
Less interference from diabetes in your social life?	261	3.5	23	29	71	61	77
Less interference from diabetes in sports and other physical activities you wish to do?	260	4.0	11	19	48	53	129

A Likert scale of 1–5 was used, where 5 is agree entirely and 1 is disagree entirely.

HbA1c, glycated hemoglobin.

Table 4.

Continuous Glucose Monitor Users Only Survey Responses to Agreement on Potential Advantages of an Artificial Pancreas System

			n
Question	Total	Average	1	2	3	4	5
Improved overall blood glucose control measured by HbA1c levels?	76	4.6	1	1	7	12	55
Fewer low glucose levels (i.e., <70 mg/dL or 4 mmol/L)?	76	4.3	0	2	9	19	44
Fewer high glucose levels (i.e., >180 mg/dL or 10 mmol/L)?	76	4.5	0	3	2	23	48
Less day-to-day swings in your blood glucose levels?	76	4.6	1	1	2	20	52
Less worry about the long-term effects of your diabetes?	75	4.6	0	1	8	15	51
Time off from thinking about your diabetes so much?	75	4.5	0	3	5	22	45
Less worry for your family members when they think about your diabetes?	75	3.5	8	10	12	19	25
Less interference from diabetes in everyday life?	76	4.3	1	4	11	13	47
A better night's sleep?	74	4.0	2	9	13	15	35
Starting the day on a “good” blood sugar number?	75	4.2	0	3	13	26	33
Less interference from diabetes in your social life?	75	3.5	7	6	23	18	21
Less interference from diabetes in sports and other physical activities you wish to do?	75	4.1	1	5	16	15	37

A Likert scale of 1–5 was used, where 5 is agree entirely and 1 is disagree entirely.

HbA1c, glycated hemoglobin.

Discussion

Our survey shows many of the potential benefits and challenges associated with AP use. Consistent with other AP research surveys, we found that respondents report a desire for optimal glycemic control, with a view to reduce risk of long-term complications.^3,4 The assumption that glycated hemoglobin (HbA1c) is the ultimate goal of diabetes management, however, is questionable. Although HbA1c is the focus of healthcare professionals and is the current gold standard measure of glycemic control (appropriate or otherwise as an average of blood glucose levels over roughly a 3-month period and failing to account for extreme lows or highs in blood glucose level), it is not always an important factor to quality of life for some people with type 1 diabetes.⁵ Many people are unaware of their own HbA1c as reported by Harwell et al.,⁶ who found that 75% of respondents were aware of what an HbA1c is and report having one in the last year; however, only 24% of those who reported having a test remembered the actual value, and the self-reported values correlated weakly with the last HbA1c value on the medical record.

Our survey results also highlights the importance that an AP device improve quality of life. The AP shows great promise in alleviating some of the burden of living with diabetes. In fact, research has shown that “time off from the demands of diabetes” is one of the associated benefits of the technology, with 91% of respondents to our survey showing significant interest in using such a device. Similarly, studies have shown that the AP can improve glucose control during the day and overnight, without increasing the risk of hypoglycemia or diabetic ketoacidosis.^7,8 Yet, participants were far less likely to indicate they would use the device if it required they provide additional information in the form of number of carbohydrates consumed or duration/intensity of exercise performed, rather than just pressing a button to announce the event, with only 62% of respondents highly interested in using such a device.

Current goals in AP research are to ensure the integrity and success of engineering features of the device to ensure safety and durability. Biomedical goals focus on time in target range and avoidance of severe hypoglycemia and hyperglycemia. Both of these are important. However, if the device ultimately does not meet the needs of those who will be using the device, success will be limited. The impact on psychosocial functioning is crucial to determine whether people will, or are able, to manage the demands of the technology in the context of their everyday lives. Something that is impractical or not fit for purpose in the holistic context of diabetes self-management will simply not be of sufficient benefit to the end user.²

There are several technologies that demonstrate engineering excellence, such as insulin pump therapy, continuous glucose monitoring, and advanced systems of self-monitoring of blood glucose such as the ACCU-CHEK^® Aviva Expert meter (Roche Diabetes Care AG, Burgdorf, Switzerland). Yet, despite the technological excellence of these devices, they have not been universally adopted or overwhelmingly successful in helping people to achieve optimal glycemic control and reduce the risk of associated long-term complications. Despite improving technical specifications of CGM devices, fewer than 10% of all persons with type 1 diabetes wear a CGM device, and almost half discontinue use within a year.⁹

There are several reported concerns about the technology that currently remain unaddressed. First, there are technical factors. In some studies, disruptions of wireless connectivity interrupted the study on average every 5 nights.⁸ In many studies, participants have worn two continuous glucose monitoring sensors in case one fails. This would not be practical for most people, as the additional of a second sensor is not feasible for everyday use. For many, the burden of adding a single sensor to use of an insulin pump is significant.¹⁰ Furthermore, the studies published to date are of relatively short duration—a few weeks at a time at most. It is not clear how well the devices will function or be tolerated for prolonged use, let alone their overall improvement in HbA1c. Finally, it is not clear how human factors, such as patient perceptions and behavioral self-regulation, will affect long-term use of these devices. Currently there is no standardized tool for psychological assessment related to AP use. Addressing this point specifically, the data from our study show that only just over half of respondents would be interested in a device that works solely overnight, even though fear of overnight hypoglycemia is a major limiting factor for current therapies for glycemic control. We had expected a much higher number to be interested in such a device. This suggests that the “technical improvements” the devices may offer are not fully addressing the underlying psychosocial concerns of persons with diabetes.

Yet, despite these concerns there is a desperate need for a device that will help minimize the burden of disease state and help facilitate improved psychosocial functioning and improved quality of life. Comments such as “I don't care, I just want it to work” reflect the need for something that will help to minimize the burden of living with type 1 diabetes.

In terms of what a successful AP would look like, 20 people said it would look like an insulin pump. This perhaps reflects the existing schema of diabetes-related medical devices but importantly suggests that people are not thinking laterally about the possibilities, as they are unduly influenced by the present and unable to see the potential opportunities for the future.

So what can we do?

The data from our study show it is imperative that we develop a standardized psychological assessment tool to be used in all future studies of AP devices. Previous studies have shown a great interest in use of an AP device, with over 85% being interested in adopting the technology once it becomes available.¹¹ But without such a tool to assess user (and parent) expectations, the likelihood of long-term use, and the psychological impact of using such a device, AP devices may suffer very slow adoption and long-term use rates similar to CGMs.

We recognize there are numerous limitations of our current study. First, the study is not necessarily representative of wider views of people with diabetes who might be eligible for or use AP technology. The study was only advertised via social media networks, so it excludes anyone who does not have access to these media. Second, we are assuming that future AP devices will be for persons with type 1 diabetes only. In reality, the device should be effective for anybody who is insulin deficient. However, the type 2 diabetes market is currently unexplored, and the challenges of usability, lived experience, and the impact on psychosocial functioning for people with type 1 and type 2 diabetes are inevitably going to be different. The differences in populations, disease etiology, obesity, age, and comorbidities mean that distinct populations will have differing needs and expectations of technology, and these will require individual assessment.

Conclusions

The views of people who would use an AP are crucial in the development of such devices to ensure they are fit for use alongside biomedical and engineering excellence. Without this, it is unlikely that an AP will be sufficiently successful to meet the needs of users and to achieve the ultimate goals of optimal glycemic control, reduced burden of diabetes self-management, and improved quality of life for people with type 1 diabetes.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Doyle

III , Huyett

, Bok Lee

, et al.: Closed-loop artificial pancreas systems: engineering the algorithms. Diabetes Care, 2014; 37:1191–1197.

Oliver

, Evan

, Hovorka

, et al.: Comment on Doyle et al. Closed-loop artificial pancreas systems: engineering the algorithms. Diabetes Care, 2014; 37:e226–e227.

Barnard

, Wysocki

, Allen

, et al.: Closing the loop overnight at home setting: psychosocial impact for adolescents with type 1 diabetes and their parents. BMJ Open Diab Res Care, 2014; 2:e000025.

Barnard

, Wysocki

, Thabit

, et al.: Psychosocial aspects of closed- and open-loop insulin delivery: closing the loop in adults with type 1 diabetes in the home setting. Diabet Med, 2015; doi: 10.1111/dme.12706.

Barnard

, Speight

, Skinner

: Quality of life and impact of continuous subcutaneous insulin infusion for children and their parents. Pract Diab Int, 2008; 25:278–283.

Harwell

, Dettori

, McDowall

, et al.: Do persons with diabetes know their (A1C) number?. Diabetes Educ, 2002; 28:99–105.

Leelarathna

, Dellweg

, Mader

, et al.: Day and night home closed-loop insulin delivery in adults with type 1 diabetes: three-center randomized crossover study. Diabetes Care, 2014; 37:1931–1937.

Thabit

, Lubina-Solomon

, Stadler

, et al.: Home use of closed-loop insulin delivery for overnight glucose control in adults with type 1 diabetes: a 4-week, multicentre, randomised crossover study. Lancet Diabetes Endocrinol, 2014; 2:701–709.

Wong

, Foster

, Maahs

, et al.: Real-time continuous glucose monitoring among participants in the T1D Exchange Clinic Registry. Diabetes Care, 2014; 37:2702–2709.

10.

Barnard

, Lloyd

, Skinner

: Systematic literature review: quality of life associated with insulin pump use in type 1 diabetes. Diabet Med, 2007; 24:607–617.

11.

Bevier

, Fuller

, et al.: Artificial pancreas (AP) clinical trial participants' acceptance of future AP technology. Diabetes Technol Ther, 2014; 16:590–595.