Abstract
Aims:
The ability of patients to achieve glycemic control depends in part on their ability to interpret and act on blood glucose (BG) results. This clinical study was conducted to determine if a simple on-meter color range indicator (CRI) could improve the ability of patients to categorize BG values into low, in-range, and high glycemic ranges.
Methods:
The clinical study was conducted in 59 subjects with type 2 diabetes (T2DM). Subjects classified 50 general, 15 before- and 15 after-meal BG values as low, in-range, or high based on their current knowledge. Subjects then interactively experienced the on-meter CRI, which showed whether alternate BG values were low, in-range, or high. After CRI interaction, subjects repeated the original scoring assessment followed by a survey exploring their awareness of glucose ranges.
Results:
Following interaction with the CRI, subjects improved their ability to categorize general, before-meal and after-meal BG results by 23.4% ± 3.0% (SEM), 14.2% ± 2.4%, and 16.1% ± 2.9%, respectively (all P < .001), into low, in-range, and high glycemic ranges. Improvement was not accompanied by an increase in time spent categorizing results. There was no correlation between subject HbA1c, test frequency, or duration of diabetes and ability to correctly classify results. Subjects agreed the CRI feature helped them easily interpret glucose values and improved their awareness of glucose ranges.
Conclusion:
A short interactive session with a meter including a CRI feature improved the ability of T2DM subjects to interpret and categorize BG values into recommended ranges.
Diabetes care groups have issued consensus recommendations on the targets for blood glucose results.1,2 The joint ADA/EASD guidelines emphasize the need to personalize glucose targets to meet individual circumstances and treatment goals. 3 The advent of self-monitoring of blood glucose (SMBG) was heralded as a significant step to enable people with diabetes to achieve improvements in glycemic control. However, there are several barriers that limit the effective use of SMBG, including lack of patient understanding about when to perform testing, how to identify and interpret glucose results, and how to respond to abnormal glucose values. 4 A recent study in 207 non-insulin-using patients with type 2 diabetes (T2DM) investigated perceptions of high blood glucose results and demonstrated patients were tolerant of higher blood glucose levels than are clinically advisable, perhaps because they do not recognize or interpret such values as being high. The authors concluded that along with performing SMBG, patients need education on appropriate blood glucose target levels and that this information needs to be reinforced over time. 5 In another recent survey of 886 US patients with T2DM, 56% did not respond regularly to out-of-range glucose readings with any self-care adjustments. 6 Furthermore, 2 UK-based qualitative studies observed that many patients found that blood glucose results were difficult to interpret and act on and that SMBG data were not routinely used to guide changes. For these patients, interpreting readings and in particular knowing how to act on high readings remained problematic. The authors suggested that the purpose of SMBG and interpretation of results should be reinforced during each clinical appointment to ensure that patients are aware how to react to high readings.7,8 It is clear from these reviews that patients with T2DM (insulin- or non-insulin-using) find individual blood glucose results difficult to interpret with respect to their own personal glycemic targets. Patient education around interpretation of blood glucose values usually occurs shortly after diagnosis with diabetes. However, it is clear that there is a need for continued reinforcement of glucose range awareness. The current study assessed the ability of subjects with T2DM to categorize individual blood glucose readings below, within, or above accepted blood glucose ranges. In addition, the study evaluated the effectiveness of a simple meter-based color range indicator (CRI) to improve the ability of these subjects to correctly interpret and categorize glucose results into these ranges. The results of this study could provide evidence for the utility of the CRI tool to reinforce health care professional (HCP) advice to their patients regarding glucose result interpretation and enable patients to follow HCP recommendations on target glucose ranges when patients conduct SMBG between office visits.
Subjects
This study was conducted at a National Health Service (NHS) clinic in the UK (Highland Diabetes Institute, Inverness, Scotland). The study was approved by the relevant internal ethics committees, and all participants provided written informed consent prior to initiation of the study. The study included participants aged ≥ 16 years with T2DM for at least 6 months. All subjects were currently self-monitoring their blood glucose. Subjects provided demographic and medical history information to the facilitator. In addition, the HCP obtained the subject’s most recent laboratory HbA1c result from the NHS SCI-DC database.
Materials and Methods
CRI Feature
Subjects experienced the CRI feature using an interactive PC-based computer simulator. The CRI tool automatically indicates whether the blood glucose result displayed on the screen is either “low” (below range), “in-range,” or “high” (above range). On the meter screen, a CRI indicator note points to a blue, green, or red bar, and the words “low,” “in-range,” or “high,” respectively, are shown within the note (Figure 1). The determination of a low, in-range, or high message depends on the range that has been set in the meter by the patient or health care professional (HCP).
Blood glucose monitor color range indicator.
However, default glucose range limits are provided preset in the meter. The default values for general glucose tests are low (<70 mg/dl), in-range (70-180 mg/dl), and high (>180 mg/dl). The default range limits for displaying a CRI note with before-meal glucose tests are low (<70 mg/dl), in-range (70-130 mg/dl), or high (>130 mg/dl). Default range limits for displaying a CRI note with after-meal glucose tests are low (<120 mg/dl), in-range (120-180 mg/dl), or high (>180 mg/dl). Default range limits were determined with respect to accepted diabetes guidelines.1 -3 Although not explored in the current study, any of the default glucose ranges can be customized by the patient or HCP.
Glucose Readings Interpretation and CRI Experience
Step 1
Subjects classified on paper 50 glucose readings as low, in-range, or high (based solely on their current knowledge) without any prior discussion of ranges. It is important to emphasize that the CRI default glycemic ranges were not disclosed to the subjects any time during the study. The study facilitator recorded the time taken for the subject to complete this exercise.
Step 2
Subjects interacted digitally with the BGM simulator to experience how the CRI tool classified a different set of 30 glucose readings. Subjects used a mouse to click on the test strip picture on the screen to visually insert a strip into the meter and then clicked on the test strip edge to simulate a 5 second blood glucose test countdown. After 5 seconds, a preprogrammed glucose reading with the appropriate CRI note (low, in-range, or high) was displayed. The subject repeated this process at their own pace to simulate 30 home glucose tests. Subjects then interactively accessed the results log (history) to see each of the 30 test readings and review the details screen, which again showed the range note associated with each result. The study facilitator recorded the time it took the subject to complete this step.
Step 3
Subjects repeated the original paper exercise in step 1 to classify the original 50 glucose readings. The study facilitator recorded the time it took the subject to complete this step.
The same 3-step process was then repeated using 15 before-meal and 15 after-meal glucose readings. Completing the entire 3-step process for general glucose readings before moving on to before-meal and then after-meal readings minimized confusion for the subjects. Once again the study facilitator recorded the time it took the subject to complete each step.
Subject Surveys
After all study procedures were completed, subjects completed questionnaires designed to demonstrate their knowledge of glucose ranges. In addition, the questionnaire explored how subjects interpret and respond to low or high glucose readings. Finally, subjects completed a brief survey on their perception of the value and utility of the CRI tool itself with respect to managing their diabetes.
Statistical Analyses
Continuous demographic variables are described as median with range (minimum to maximum) or mean and standard deviation (SD). Categorical demographic variables are described as percentages within categories. Test score changes were calculated as 100 × (posttraining score – pretraining score) / (pretraining score), expressed as the % change from baseline ± standard error of the mean (SEM). The null hypothesis “H0: Pretraining score = posttraining score” was tested using a paired t test with significance level α = .05. Correlations with HbA1c were assessed using the Pearson correlation coefficient and were deemed significant with a P value < .05. Minitab 16.1.1 was used for all analyses.
Results
Subjects
Baseline characteristics of the 59 subjects with T2DM who completed the study are shown in Table 1. The mean age of subjects was 65 years with a mean and median duration of SMBG use of 11.7 and 9.8 years, respectively. Of subjects, 66.1% reported they perform SMBG at least 2 times per day. Of the study population, 28.8% were currently on insulin therapy alone, 45.8% on insulin and oral antidiabetic drugs (OADs), and 18.6% on OADs alone. The median HbA1c across all 59 subjects was 8.1% (5.9 to 13.9%).
Baseline Patient Demographics and Medical History.
Ability of Subjects to Classify Readings Into Appropriate Glucose Ranges
During step 1, on average study subjects correctly classified 34.7 of the 50 general readings into appropriate ranges. However, after exposure to the CRI tool (step 2), the average correct number of responses increased to 42.8, representing a statistically significant improvement against baseline across all subjects of 23.4% ± 3.0% (SEM) (P < .001). Subjects took 2.4 minutes on average to classify all 50 general readings during step 1, and 2.1 minutes after CRI interaction. The average time spent interacting with the CRI tool itself (during step 2) was less than 6 minutes, with a range of 1-14 minutes across subjects. Individual subject improvements varied considerably with 30%, 20%, and 10% of subjects improving by 33%, 47%, and 67% or greater, respectively, after experiencing the CRI tool (Figure 2A).
Change in glucose result classification scores. (A) General range, (B) before-meal range, (C) after-meal range. Each bar represents an individual subject.
During step 1, on average study subjects correctly classified 9.7 and 9.9 of 15 before- and 15 after-meal readings into appropriate ranges. After exposure to the CRI tool (step 2), the average correct number of responses increased to 11.0 and 11.5, respectively, representing a statistically significant improvement against baseline across all subjects of 14.2% ± 2.4% (P < .001) and 16.1% ± 2.9% (P < .001). On average, subjects took 45 seconds and 44 seconds respectively to classify 15 before- and 15 after-meal readings during step 1. The time spent to classify before- and after-meal values during step 3 (47 and 51 seconds, respectively) was not statistically different from step 1. Individual subjects varied in their ability to classify mealtime readings after CRI interaction with 30% of subjects improving their baseline before- or after-meal scores by 22%, and 30% or greater (Figures 2B, 2C). The average time spent interacting with the CRI tool itself for mealtime readings was 2.6 minutes (range 1 to 6 minutes) for before-meal and 2.3 minutes (range 1 to 6 minutes) for after-meal readings.
Range Classification and Clinical Measures
The HbA1c of subjects varied between 5.9% to 13.9%. There was no correlation between their HbA1c and the ability of subjects to classify 50 glucose readings into appropriate glucose ranges. There was also no correlation between the subject SMBG frequency or their duration of diabetes and their ability to correctly classify glucose readings.
Subject Surveys
Subject Awareness of Glucose Ranges
Of subjects, 64.4% (38) confirmed they had previously discussed glucose ranges with their HCP (Figure 3A). Of these 38 subjects, 47.4% responded that their HCP mentioned glucose ranges at every scheduled visit, and 23.7% that their HCP mentioned it only sometimes (Figure 3B). Furthermore, 31.6% of these 38 subjects stated their HCP had previously given verbal and written glucose ranges, while 42.1% reported that they only received verbal information (Figure 3C).
Subject awareness on glucose ranges. (A) HCP ranges discussion, (B) frequency of HCP range discussion, (C) type of HCP ranges feedback.
Subject Perceptions of High or Low Glucose Readings
When subjects were asked “Is there a low glucose value you would prefer to stay above during routine SMBG?,” 53/59 specified a low limit to stay above. Of these 53 subjects, 50 (94%) responded that they would prefer to be above at least 3.9 mmol/l with 28 (53%) preferring to always be above at least 5.0 mmol/l (Figure 4A). When asked “Is there a high glucose value you would prefer to stay below during routine SMBG?”, of the 54 subjects who specified a high limit to stay below, 16 (30%) responded they would prefer to stay below 10.0 mmol/l, with the remaining 38 (70%) preferring to stay below higher target limits (Figure 4C). When subjects were asked “Is there a low glucose reading at which you would take action (to raise blood glucose)?,” of the 59 subjects, 58 (98%) responded they would act below 3.9 mmol/l (Figure 4B). When the subjects were asked the same question for high readings, 30 (51%) responded they would not take any action at any high glucose reading, with 29 (49%) responding they would act above 10.0 mmol/l (Figure 4D).
Subject perceptions of low and high glucose readings. (A) Perceptions of low readings to stay above during routine SMBG. (B) Level at which subjects would take action for a low glucose result. (C) Perceptions of high readings to stay below during routine SMBG. Subject perceptions of actions associated with low and high glucose readings (D) Level at which subjects would take action for a high glucose result. The numbers above each bar denote the number of individual subjects who selected this glucose range or value.
Subject Perceptions of the CRI Tool
A survey designed to explore subject perceptions of the CRI feature showed that 96% agreed that the CRI made them more aware of their target ranges and could help them quickly (92%) and easily (90%) interpret their blood glucose readings. Furthermore, 88% of subjects felt that the CRI could help them (1) easily report to their HCP when readings were in or out of range and (2) talk to their HCP about low and high glucose limits that were right for them (86%) to help them reach their blood sugar goals (82%) (Table 2)
Patient Responses to Survey Statements.
Favorable responses are defined as responses of strongly agree or agree on a 5-point scale (5 = strongly agree, 4 = agree, 3 = neither agree nor disagree, 2 = disagree, 1 = strongly disagree). All favorable responses are statistically significant (P < .05).
Discussion
A brief digital interactive exposure to a novel meter-based CRI providing automatic on-screen glucose range notes was demonstrated to significantly improve the ability of patients with T2DM to classify BG readings into accepted low, in-range, or high glycemic ranges. Such improvements were not at the expense of taking more time to evaluate blood glucose results. These findings have positive implications for this CRI tool with respect to improving a patient’s ability, particularly after sustained “real-world” exposure, to follow their HCPs recommendations to maintain blood glucose readings within target ranges agreed on with their HCP.
The ADAG study group recently reported the recommended average blood glucose needed to achieve a variety of HbA1c targets. 9 With regard to these recommendations, Klonoff 10 stated that the “practice of SMBG has become more important than ever because for the first time there is now excellent evidence presenting which levels of blood glucose are needed to achieve intended levels of mean glycemia.” Our findings support the utility of a meter-based CRI to help HCPs reinforce advice about individual SMBG targets discussed during consultations. In the present study, there was an improvement in the ability of subjects to recognize if general blood glucose results were below, within, or above a target glucose range even after a relatively brief exposure to the CRI and without any formal discussion of glucose ranges. Similar results were also seen using representative before- and after-meal glucose results.
ADA 11 recommendations state “optimal use of SMBG requires proper review and interpretation of SMBG data, both by the patient and provider.” In the present study, a range of baseline capability to interpret individual glucose readings as low, in-range or high was seen. This individual variation is not surprising. Evans et al 12 observed a lack of knowledge surrounding SMBG among non-insulin-treated patients, with more than one-third of 207 patients reporting that they took no action in response to SMBG results. Over half of those who reported taking no action stated they did not know what to do. Similarly, in the present study we observed that although most subjects preferred to maintain high glucose readings below 180 mg/dl, many stated that they would take no action at any high glucose reading. In contrast, the majority of subjects strived to maintain low glucose values above 70 mg/dl, with most subjects stating they would only take action below 70 mg/dl. Evans et al 12 surveyed only non-insulin-using patients with T2DM whereas our study had a high proportion of subjects either on insulin with oral agents or insulin injections alone. It is possible that these subjects may have been considering adjusting insulin to address high glucose readings. Despite this, our data suggest a similar response to Evans et al 12 with the majority taking no action at any high glucose reading.
Difficulties interpreting blood glucose results and comprehension of acceptable glucose ranges are not confined solely to patients with T2DM. Rankin et al 13 performed in-depth interviews with 30 patients with type 1 diabetes and whilst blood glucose targets were considered essential to achieving tight glycemic control, little attention had been paid to how they were operationalized within everyday life. In follow-up interviews, patients frequently struggled to recall precisely their recommended glucose targets and had difficulties attaining or sustaining recommended SMBG readings. It is possible that the issue highlighted by Rankin et al, 13 whereby patients struggled to recall glucose targets, could have been minimized by a customizable CRI. It is important to note that at no point in the present study were subjects alerted to the actual range settings used by the CRI tool. Rather, subjects simply experienced how the CRI tool interpreted individual readings, and this experience appeared to enable subjects to deduce where glucose results may fall within a target range. It is also important to recognize that subjects made encouraging improvements despite only having a short experience with the CRI tool. The variability of improvement (and lack of improvement in a few instances) across study subjects also suggests that the CRI experience and glucose results interpretation exercise was set at an appropriate level of difficulty, enabling individual differences in aptitude for improvement to surface. If the CRI experience had been too extensive and the results interpretation exercise less challenging, then we may have seen all subjects making substantial improvements (or the reverse if the CRI experience had been insufficient to change awareness and the results exercise too challenging).
It seems logical that subjects with better glycemic control might be better able to interpret and classify blood glucose readings into appropriate ranges. Notably, Garg et al14,15 considered the question “What should the percentage of in-range values be to have a good A1c value?” and reported a strong correlation between HbA1c and the percentage of glucose readings within, above, and below target range. However, we observed no correlation between HbA1c and ability of subjects to classify results into appropriate ranges. This lack of association is perhaps not surprising given that both Evans et al 12 and the current study demonstrate that subjects seldom take action to control high glucose readings, even if they correctly interpret them as high. Supporting this idea, a recent review by the AADE summarized the extent to which both type 1 and type 2 patients on insulin therapy have difficulty determining appropriate insulin doses, fail to take HCP prescribed doses or intentionally skip insulin doses. 16 These behaviors clearly may diminish the potential to maintain glycemic control in response to BG values. The CRI tool could encourage patients to bring out of range BG trends to the attention of their HCP to obtain advice on appropriate action.
In the subjects participating in the current study, there was no correlation between SMBG test frequency or duration of diabetes and the ability to classify glucose results. Such findings suggest that the ability to interpret data does not necessarily translate into action or effective action. In addition, Wang et al 17 reported that when 555 non-insulin-using type 2 patients considered their own blood glucose too high, 50% reported doing nothing and only 26% resolved the problem. When blood glucose was low, 68% of these patients also did nothing and only 30% reported achieving a normal range within 30 minutes.
Recent reviews18,19 reveal an increase in research focusing on problem solving and diabetes knowledge as modifiable skills for diabetes self-management, and a recent consensus report 20 on SMBG reinforced the need for patient education on how to respond to the data for SMBG to be effective. Heller et al 21 also identified a lack of education on how to interpret and use data from SMBG. Other obstacles to blood glucose monitoring such as inadequate numeracy or health literacy are seen as contributing factors.22,23 The intrinsic value of color to enhance a patient’s understanding of blood glucose readings has been advanced by Spollett, 24 who suggested a traffic-light-type color-coded system with a first step being agreement on a target glucose range. Spollett explained that clustering of any 1 color at a particular time of day is not a judgment of patients’ efforts, but rather indicates that they need to review that time frame for possible issues and action steps to achieve targets. The utility of a color record tool in association with diabetes education to improve HbA1c has been demonstrated recently by Nishimura et al, 25 who hypothesized that color used in SMBG is helpful for patients to recognize and act on their glucose levels and demonstrated a lowering of HbA1c after 24 weeks in subjects utilizing a simple color record tool to highlight individual low or high results. A recent review by Kumah-Crystal and Mulvaney 26 indicated a need to improve and expand support for people with diabetes in their daily use of glucose values to maintain and improve glycemic control. Subject responses in our current study support the utility of a meter-based CRI to facilitate improved awareness of glucose ranges and provide easier interpretation of readings and the potential to make it easier to report in or out of range results to their HCP.
Summary and Conclusions
Evidence suggests that patients with diabetes and in particular patients with T2DM have minimal ongoing education on blood glucose data interpretation, have minimal recall of HCP target ranges, and often do not know when (or how) to respond to individual glucose results. To help address some of these issues, we have provided evidence that a meter-based CRI tool was associated with improved patient interpretation of glucose readings, enabling patients to better categorize readings as low, in-range, or high, with regard to accepted glucose ranges. Longer term exposure to the CRI tool (particularly when the CRI is individualized) may enable patients to better adhere to HCP-suggested blood glucose targets and alert patients when to communicate with their HCP.
Footnotes
Acknowledgements
The authors would like to thank Nichola Purvis, Krisna Koria, and Lorna Stewart for support of the study conduct.
Abbreviations
BG, blood glucose; BGM, blood glucose monitor; CRI, color range indicator; HbA1c, hemoglobin A1c; HCP, health care professional; NHS, National Health Service; OAD, oral antidiabetic drug; SCI-DC, Scottish Care Information—Diabetes Collaboration; SD, standard deviation; SMBG, self-monitoring of blood glucose.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MG and GW are employees of LifeScan Scotland, Ltd. BLL and LBK are employees of LifeScan Inc.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by LifeScan Inc. The authors received editorial assistance from Emphatic, which was funded by LifeScan Inc.