Usability Evaluation of a Blood Glucose Monitoring System With a Spill-Resistant Vial, Easier Strip Handling, and Connectivity to a Mobile App

Abstract

Background:

Proper management of diabetes mellitus requires regular self-monitoring of blood glucose (SMBG). This research evaluated the usability of the Accu-Chek® Guide Meter that includes a spill-resistant vial, easier strip handling, and wireless connectivity to a mobile app.

Methods:

A total of 197 participants were allowed to experience typical blood glucose testing tasks on the Accu-Chek Guide Meter, review data such as last result, patterns, and target percentage on the meter and on the mobile app, and then evaluate their experience through a human factors usability survey. Participants used a 6-point agree/disagree scale to rate 34 market statement questions.

Results:

The results of a Pearson chi-square proportions test for each of the 34 market statement questions showed a significant difference (P < .0001) between the disagree responses (1-3) and agree responses (4-6). An overwhelming majority of participants found all aspects of the system, including the test strips, strip vial, and data analysis on the meter and the mobile app, to be a good fit for their lifestyle and to provide a better testing experience.

Conclusions:

This study found superior usability of the new meter system over the participants’ current meters in both the United States and France.

The proper care of diabetes mellitus is a burden for patients, but also puts a financial burden on society. According to the Centers for Disease Control and Prevention, as of 2010 diabetes affects 25.8 million people across all ages in the United States, including both diagnosed and undiagnosed diabetes.¹ Diabetes is the seventh leading cause of death in the United States, partially due to complications seen in the macro- and microvascular bed, and the costs of diagnosed diabetes exceeded $245 billion in 2012.^1,2 Unfortunately, these numbers are only going to increase in the upcoming decades; 1 group studied the trend in the prevalence of diabetes over 20 years and found that the prevalence of diagnosed diabetes increased from 5.1% in 1988-1994 to 7.1% in 1999-2004 to 8.4% in 2005-2010.³ Given these astonishing statistics and trends, proper care and management of this disease are essential to mitigate the diabetes epidemic.

The widespread use of in-home blood glucose (bG) meters over the past 20 years has revolutionized the way that people with diabetes are able to care for and manage their disease.⁴ Structured self-monitoring using an in-home bG meter not only provides a way for the patient to track his/her bG, but allows the patient to manage the disease by observing the metabolic fluctuations induced by food intake, physical activity, stress or fever, or the development of hypoglycemia.⁴ The frequency and timing of self-monitoring of bG is different for each individual based on the unique manifestation of the disease, doctor recommendation, and the willingness of the patient to use the meter.⁵ Despite an individualized regimen of bG testing, sometimes patients do not test their bG as frequently as recommended because of the inconvenience and discomfort of testing or because of testing errors despite the proven accuracy of the meter.⁴ In addition to regular testing, patients are encouraged to measure bG when there is a suspicion of a hypoglycemia, since data suggest the prevalence of hypoglycemia results is a major cardiovascular risk.⁶ However, testing under hypoglycemic conditions sometimes results in testing errors since hand-eye coordination is often diminished.

While inconvenience and discomfort are an inherent part of the management of the disease, reduction in the frequency of errors can be addressed. While technical advances have greatly reduced mechanical errors in the accuracy of the meters, research indicates that user error, not mechanical error, is the biggest source of inaccurate bG readings.^4,7 In 1 study by Colagiuri et al,⁸ 62% of patients made at least 1 “clinically significant” error, with “faulty technique” being cited as the cause for most errors. Another study found that roughly 75% of the errors made could be traced back to user error; the most common source of error in this study was correct application of blood to the test strip.⁹ While these studies were conducted on older generations of meters than the one being tested, user errors still exist today.¹⁰ Increased usability of the bG meter and test strips could play a role in patients’ willingness and ability to accurately test according to the doctor recommended regime.¹¹ Several known handling problems in meter system designs resulting in errors in bG testing are identified in Table 1.

Table 1.

Frequent Handling Problems in Glucose Self-Monitoring.

Action	Handling problem	Potential error
Get the strip from the vial container	• Strips stick together • Strips spill from container	• Cost of wasted strips12 • Reduced frequency of testing due to inconvenience, discretion concerns,12-14 reduced quality of life • Inability to perform required action, especially with symptoms of hypoglycemia14 • Inaccurate result, possibly wrong clinical decision
Insert strip and apply the blood sample to the strip	• Blood drop too small or not placed correctly10 • Can’t see strip/strip port due to low lighting • Desire to be discrete13
Read results on meter	• Previous reading not visible • Target bG levels are not set/available15	• Misinterpretation of results • Lack of educated decision making guidance13

In the past, usability was not necessarily a primary goal when developing and testing new medical devices, but the publication of “Medical Devices—Application of Usability Engineering to Medical Devices”¹⁶ has led to a new push to develop medical devices that are both safe and easy to use. Usability engineering, a technique that relies on user input in the early stages of development to improve the design of a product, is especially important in the field of diabetes management because diabetes patients have a greater role in self-management and decision making than in most other medical conditions.¹⁷ After accuracy, ease of use was the second most important attribute to be considered when purchasing a bG meter.⁴ Usability testing prior to releasing a new bG meter on the market can help clarify user constraints, wants, and capabilities, as well as design out potential errors and identify important user needs, thus enabling patients to follow their doctor-recommended testing regimen in the easiest and most accurate way possible.¹⁷ In addition to ease of use and reduction of user error relating to testing bG, incorporating other features to allow a user to keep track of results, provide recommended insulin doses, and identify overall trends in bG readings as well as connectivity to a smartphone could also encourage consistent use of the meter for better diabetes management.¹⁸

Methods

Louisiana State University in conjunction with Roche Diabetes Care, Inc conducted a human factors usability claims validation study of the Accu-Chek® Guide Meter System designed to address the common handling problems previously listed in Table 1. The study focused on several new features of the system that are intended to reduce the occurrences of common handling problems. For example, the design of the strip vial prevents strips from spilling, and the strip has a large dosing target. The meter itself includes an ejection button for strip removal as well as a light illuminating the strip port. The pattern detection and data management on both the meter and via wireless connectivity to a smart phone allow the user to view current and past results, trends in bG readings, and obtain insulin bolus advice. Images of the meter, strip vial, and smartphone app are shown in Figure 1.

Figure 1.

Accu-Chek Guide Meter, spill-resistant strip vial, and smartphone application.

The objective of this study was to evaluate the participants’ perceived usability (level of agreement/disagreement) of the meter after performing some typical tasks that include performing bG testing, use of the pattern detection feature, and the review of data on both the meter and with the mobile app. The study was broken down into the 3 phases shown in Figure 2. Following each phase, participants were asked to complete a series of usability statement questions using a 6-point agree/disagree scale.

Figure 2.

Phases of the study.

Participants

A total of 197 participants from the United States and France met the following basic inclusion criteria:

1. Be insulin dependent and inject insulin at least 1 time per day

2. Perform all bG meter tasks on their own (no extreme visual or dexterity impairments)

3. Test bG at least 2 times a day

Participants were roughly evenly divided by gender. The age range was 18-69 years old with approximately 50% of the total participants falling into the 35-59 years age group. More than 70% of the participants were considered technologically adept (defined as using a smartphone and downloading at least 1 application in the past 6 months and using apps at least 1 time per week). Table 2 shows the breakdown by current meter brand being used by participants with 76% of users having used their current meter for 5 years or less.

Table 2.

Current Meter Used by Participant.

Current meter	Abbott	Bayer	LifeScan	Roche	Total
N (%)	40 (21)	27 (13)	73 (37)	57 (29)	197

Statistical Analysis

A 6-point rating scale was used where quite agree is the positive side of scale with a value of 6 and substantially disagree is the negative side of scale with a value of 1. The ratings were split into binary responses with quite agree, moderately agree, and perhaps agree representing a positive response and substantially disagree, moderately disagree, and perhaps disagree representing a negative response. These terms have been psychometrically determined to be equal distance apart from one another.¹⁹ In addition, the lack of a middle point ensured that users chose a positive or negative response. This allowed for the positive and negative sides of the scale to be collapsed and compared for each statement. This approach determined the proportion of the sample that agreed versus disagreed with each statement. A Pearson chi-square proportions test was used to determine if the proportions of the responses to the 2 sides of the scale are significantly different from one another. An α of .05 was assumed for all testing.

Results

The first step in performing the proportions test was to combine the data from both the US and the French testing locations and group the data into agree and disagree binary responses. The proportions test only evaluates the number of agree and disagree statements and not the level of those measures. The total number of participants (N) was 105 from the United States and 92 from France for a total of 197 for most questions, except where a participant failed to mark a response. Results of a Pearson chi-square proportions test showed a significant difference (P < .0001) between the disagree responses (1-3) and agree responses (4-6) for all statements in each phase (Figures 3 and 4). These results show that significantly more people rated each statement on the agree side of the scale than the disagree side of the scale. Above is a summary of the proportions testing broken down by statement showing the percentage of answers for each response category.

Figure 3.

Phase I proportions analysis.

Figure 4.

Phases II and III proportions analysis.

Discussion

Proportions test results showed that for all statements, regardless of country, significantly more participants agreed with the usability statement questions than disagreed. A total of 95% or more of participants agreed with 21 out of a possible 34 statements. For all 11 statements related to guidance and advice in Phase III (Q6a-c), more than 95% of the participants agreed with the statements. This indicates that guidance and advice available on both the meter and the mobile app is of vital importance and participants found the guidance and advice features to be much attuned to their needs. Only 4 statements out of 34 resulted in less than 90% of the participants agreeing with the statement questions; however, the proportion who agreed was still significantly more than those who disagreed. Incidentally, these 4 statements were the only statements comparing the meter system to the participants’ current system (applying blood, removing a strip from the container, and performing quick/easy bG test). These statements also appeared during the Phase I of the session, which means participants only had limited interaction (usually less than 20 minutes in duration and performed 3 bG tests) with the new device before responding to the statements. When one considers that the participants have been using their current meter potentially for many years and were exposed to the meter only for less than an hour, achieving 84-90% agreement on these 4 statements speaks highly of the usability of the meter. Overall, these results indicate that the participants significantly agreed with all of the usability statement questions made by the meter.

Additional grouping was done for 2 statements in question 1 to determine how many users found that the strip made it easier and faster to correctly place a small blood drop on the end of the strip. Results showed that 93% of users (184 out of 197) agreed with both statements 1A and 1B. In addition, 2 statements in question 3 were combined to see how many users found that checking bG levels with the meter was easier and faster than with the participant’s current system. Results of this grouping showed that 83% (163 out of 197) of participants agreed with both statements 3B and 3C. In each case, the proportion who agreed with both statements was significantly different than those who disagreed with one or both statements.

Last, analysis of variance (ANOVA) was used to determine if current meter brand resulted in any significant difference in response to any questions. Results of this analysis found that there was no significant difference in any question for either France or the United States based on the participants’ current brand of meter. This result indicates that participants did not rate any question differently based on their current meter brand.

Conclusion

A discrete, comfortable, and easy-to-use bG self-management system is the key to successful diabetes management. Designing meters that take usability into consideration can help people with diabetes better manage the disease. Trying to address some of the most common handling problems, the meter was developed and then tested by current diabetes patients for usability. User responses to the meter were extremely positive when reviewed using proportions testing. An overwhelming majority of participants found the meter, including the test strips and strip vial, to be a good fit for their lifestyle and provides a better testing experience compared to their current systems. Participants were able to successfully use the meter and the features with very little training or exposure to the meter regardless of the brand of their current meter. The pattern detection and guidance features available on the meter and the mobile app were also very well received by most of the participants indicating that these state-of-the-art features are a necessary step in effective diabetes care management in the future. The meter, Accu-Chek Guide, was found to have superior usability over the participants’ current meters in both the United States and France.