Background
Correct knowledge of diabetes management targets is associated with better glycemic control and improved diabetes self-care [
1‐
4]. However, past studies estimate that as few as 25% of people with diabetes can accurately describe the meaning of the hemoglobin A1c value (HbA1c) or recall their most recent value [
2,
5]. While this may not seem surprising given the conceptual complexity of the HbA1c value (e.g., expressed as a percentage, non-intuitive goal range), even simpler assessments of diabetes control appear difficult [
6‐
8]. In two prior studies, individuals with poor glycemic control were asked to more generally describe their current diabetes control using Likert scales with qualitative descriptors. Many of these individuals, particularly those with low health literacy, erroneously described their diabetes as well-controlled in spite of average HbA1c values over 9% (11.7 mmol/L) [
7,
8]. While these findings may simply indicate a lack of patient understanding of the HbA1c, they may also reflect differences in the ways patients and providers conceptualize and gauge diabetes control. Though providers may expect patients to also use the HbA1c value, the rubric used by patients to assess control may be different and remains incompletely understood.
In several established models of health behaviors and outcomes, individuals’ awareness and assessment of their current disease status are important predictors of behavior change or outcomes. Examples include perceived “disease severity,” cited in the Health Belief Model, the role of “consciousness raising” in the Transtheoretical Model of Health Behavior Change
, where knowledge or information can contribute to a shift from the pre-contemplation stage to the contemplation stage, and the “informed, activated patient” in Wagner’s Chronic Care Model [
9,
10]. Still, little remains known about the factors that contribute to these disease-related assessments for patients with diabetes. Better knowledge of the factors influencing patients’ evaluations of their diabetes may enable providers to communicate more effectively regarding diabetes management targets and current levels of diabetes control.
We conducted semi-structured interviews with patients with diabetes to address the following research questions: 1) What information do patients use to assess their current level of diabetes control? and 2) What are patient-perceived barriers to understanding the HbA1c value?
Results
Between June 2013 and December 2013, approximately 240 individuals were approached regarding participation. Of those approached, 83 reported having diabetes, and 25 individuals agreed to participate. Of this 25, 68% were women and 84% were Black. There was a wide range of educational attainment, the mean duration of diabetes was 11 years, 36% of participants self-reported a diabetes-related health complication, and 40% reported current treatment with insulin (Table
1).
Table 1
Participant characteristics and diabetes history
Age (mean years ± SDa) | | 57 ± 13 |
Gender |
Female | 17 | 68 |
Ethnicity |
Hispanic | 1 | 4 |
Race |
Black | 21 | 84 |
White | 1 | 4 |
Multipleb | 3 | 12 |
Education |
Less than High School | 3 | 12 |
High School or GEDc | 13 | 52 |
Some College/Technical School | 6 | 24 |
College or beyond | 3 | 12 |
Years since diabetes diagnosis (mean ± SD) | | 11 ± 7 |
Experienced a diabetes-related complication (Yes) | 9 | 36 |
Diabetes Treatment |
Oral medications only | 12 | 48 |
Insulin | 6 | 24 |
Oral medications & Insulin | 4 | 16 |
Diet only | 3 | 12 |
Factors influencing participants’ assessments of diabetes control
Participants’ assessments of current control fell into the following thematic domains: 1) perceived self-efficacy and adherence to self-management; 2) the types and amount of medications taken; 3) the presence or absence of symptoms attributed to diabetes; 4) numerical data, both the HbA1c and self-monitoring of blood glucose (SMBG); and 5) connections between these domains.
1) Perceived self-efficacy and adherence to self-management
Of the 25 participants, seven referred to diabetes control in terms of self-efficacy, expressing confidence in their abilities to complete self-care and prioritize their health. For 17 participants, diabetes control was intrinsically linked with their perceived adherence to diabetes self-management behaviors. For the majority of participants (
n = 17), self-care activities used in assessing diabetes control centered on maintaining a healthy diet, exercising, and weight management. In addition, four participants cited routine contact with providers as part of self-care used to assess their current diabetes control.
I’m on top of my job…I’m on top of doing what I’m supposed to do to maintain this thing here. (P4)
[Referring to poor control] Not being disciplined, not setting my priorities and my priority is my body. (P9)
If I keep this excess weight off me. That’s how I’ll know [about level of diabetes control]. (P25)
I go to the doctor. Make sure that I get examinations for my eyes and for my feet, and come into my doctor regularly. (P18)
2) The type and amount of medication taken
Perceived medication adherence, as well as the number and type of medications taken, affected ten of the participants’ perceptions of their current control. The ability to stop a medication (based on a physician’s recommendation) and not requiring insulin therapy were both considered markers of good control.
When you’re not taking your medicine or doing what the doctors told you, you could tell by your health digressing. Like, more and more complications could arise or whatnot. (P13)
It’s well under control because I got a letter from my doctor telling me that my diabetes is well under control and if I wanted to stop the medication, I could. (P5)
I took insulin three times a day. And my doctor told me I’m doing good because I haven’t been back on insulin in ten years. (P8)
3) Presence or absence of symptoms attributed to diabetes
Over half of the participants (n = 14) cited the presence of diabetes-related symptoms as an indicator of “bad” diabetes control. While eight participants mentioned specific symptoms often ascribed to hyperglycemia (e.g., polyuria, polydipsia, blurry vision), others mentioned symptoms not as readily attributable to diabetes, such as pain (n = 3) or general feeling of low energy and malaise (n = 9).
Because, of course because you feel better, for one thing. And like I said, some of the things like I was doing, running to the bathroom frequently. (P15)
Because sometimes it be real high cause you get headaches and a lot of pain. Your bones ache and stuff and you’re in pain. (P21)
Things in your body that say, hey, it’s not right. (P18)
How do I tell if I’m in good control? When I don’t feel tired. (P7)
4) Numerical data
Of the 25 participants, 16 referred to the HbA1c without prompting (i.e., prior to any mention of the HbA1c by the interviewer) as information used to assess diabetes control. However, seven patients did not refer to the value correctly by name (e.g., A1, AC1 U). Besides the current HbA1c, eight participants stated that improvements in the HbA1c over time were an indicator of good diabetes control. Nearly all participants (n = 21) also mentioned the use of SMBG as a means of assessing current diabetes control.
Numbers—like when you take your sugar, hopefully all your numbers are where they’re supposed to be. (P1)
I have to do finger pricks. I take my—got machine where I take my diabetes, and that tells me what amount the diabetes is. (P16)
5) Connections between domains
There were several commonly made connections between domains. Adherence to diabetes self-management or medications, together with numerical data (SMBG or HbA1c), was a combination that 11 participants used to assess their diabetes control. Another combination used by 13 participants to assess current control was numerical data (SMBG or HbA1c) along with the presence or absence of diabetes-attributed symptoms.
I try to keep my weight down, but I’m picking up weight now. So I’ll be looking any day when I come to the doctor that she tells me that the count [referring to the HbA1c] is up again because I’m gaining weight. And I know that. (P8)
Doing well, yeah, by my finger sticks. And even when I go to the doctor, they check the blood through the arm. That’s when they told me my A1c was up. But now it’s down because I’ve been taking the medicine twice. (P7)
It’s mainly the hemoglobin A1c and just how I feel. (P22)
Exploring understanding of the HbA1c value
Although nearly all participants (n = 22) had at least heard of the HbA1c in the past, and the majority mentioned the HbA1c without prompting (n = 16), understanding of the value varied greatly. When asked what the value meant, six participants described what the test measured (e.g., an average over a period of time), ten participants described their general interpretation (both accurately and inaccurately) of the value (i.e., goal to be lower, target range), three participants described a more personal importance of the value, and three participants had just heard of the test before but could not provide further description. After being given the basic description of the HbA1c, seven participants reported that this information changed their previous understanding in some way.
How much sugar was in my body for the last three months. The HbA1c measures for the last three months. And it’s supposed to be seven or below. But I keep mine at 6.2, once in a while at 6.4. Once it goes 6.4, I get back on my strict diet to knock it down to 6.2. I don’t like for it to pass 6.2. (P24)
No, I don’t know the meaning. All I know is it’s got to be low instead of high. (P21)
Like you could do your sugars every day and you could see it, but that’s the main number that shows that all of them numbers. (P3)
It means lot, it means that this is something that can affect my health if I don’t stay on top of it. (P10)
I know it is life saving for me and like it helps like it’s a super guideline. If it is off course, then my doctor can correct it through the medication. (P11)
Several participants (n = 3) thought of the HbA1c as a sort of detector of non-adherence to diabetes self-management.
And so, that level [referring to the HbA1c] let me know I was cheating, and I have to just get back on track. (P20)
Because that’s the only way that you can really find out if a person is really doing what they supposed to be doing as far as diabetes is concerned. (P4)
Discussion
While the HbA1c was an important contributor to participants’ evaluations of their diabetes control, participants also considered other information, including perceived self-efficacy and adherence to self-care activities, the types and amounts of medications used, the presence or absence of symptoms attributed to diabetes, and SMBG.
The importance participants placed on diabetes self-efficacy and self-care activities is well-founded and supported by existing literature [
12,
13]. Provider awareness of the roles that self-efficacy and self-care activities play in patients’ model of diabetes control is vital. By focusing on building self-efficacy, and supporting and reinforcing self-care in their communications with patients, providers’ recommendations and feedback can be more in line with patients’ perspectives and priorities, hopefully increasing the effectiveness of these communications [
14].
Medications, particularly the use of insulin, were linked to many participants’ perceived diabetes control. The belief that more medications or insulin use is an indicator of worse or “end-stage” disease is established in the current literature and is a common barrier to medication intensification, particularly insulin initiation [
15,
16]. Dispelling the idea that insulin use is an indicator of current diabetes control and addressing beliefs regarding the number of medications taken could shift patients’ perceptions regarding diabetes control. Re-framing the types of pharmacologic treatments prescribed not as failures of management, but rather as results of individual physiologic differences, could have a beneficial impact on patient-provider communication, patient self-efficacy, and adherence to these prescribed medications.
Associations of patient-reported symptoms and symptom severity with diabetes outcomes, including patients’ self-rated health status and measured glycemic control, have been noted in past work [
17,
18]. Although symptoms can be an indicator of poor glycemic control, two potential issues complicate patients’ use of this metric to gauge their current diabetes control. First, patients and providers may differentially attribute certain symptoms to diabetes. Patients may attribute symptoms such as general malaise and fatigue to poorly controlled diabetes, even though these are symptoms providers may not immediately associate with this disease [
19,
20]. Notably, malaise and fatigue may actually be symptoms of depression, a condition that disproportionately affects patients with diabetes, remains underdiagnosed in this population, and may contribute to worse diabetes outcomes [
21]. Second, the absence of symptoms may provide false reassurance to patients regarding their current diabetes control. Many patients with blood glucose levels well above goal do not experience any symptoms in spite of their increased risk for future diabetes-related complications. Providers should emphasize this point to patients in their communications about diabetes control.
Numerical information, either SMBG or the HbA1c, was part of many participants’ assessment of diabetes control. The frequent mention of SMBG as an important source of information is noteworthy given mixed evidence regarding the reported clinical value of ongoing SMBG in patients not on insulin (only 40% of interviewed participants reported insulin use) [
22]. The importance given to the HbA1c in assessing control and management was somewhat surprising given the varying levels of actual understanding of the value’s meaning. Given past evidence supporting the value of an accurate understanding of disease management targets, like the HbA1c, further efforts are needed to improve the way care providers present this information to patients with diabetes [
2‐
4]. Improving providers’ communication of this value could address the current barriers to HbA1c understanding and, hopefully, increase the HbA1c’s informational value for patients with diabetes. Further work is needed to identify optimal approaches for communicating information on glycemic control to patients. Towards this goal, in an in-progress mixed-methods study, we elicited input from participants with diabetes on a variety of visual formats for presenting the HbA1c value (e.g., color-based scales, depictions of distance from goal). Based on qualitative analysis of participants’ input, two formats were chosen for testing against standard presentation of the HbA1c in a three-arm randomized, controlled trial of patients with diabetes to assess their impact on patients’ assessments of their current diabetes control.
The study has several key limitations. First, the population is small and demographically homogenous, limiting the generalizability of the findings to other populations. Of note, the race/ethnic demographics of the participants do reflect that of West Philadelphia, a predominantly Black neighborhood. Because demographic information was not collected from eligible individuals who declined participation, differential participation by race or educational attainment cannot be assessed. Second, while the investigator AG’s supervision and coding of the initial interviews may have introduced potential bias, it was necessary for the interviewer’s (KM) training and unavoidable given limited study staff. We feel that the use of a structured interview script, coding by two independent individuals, and input from the broader study team during the analysis process, helped to mitigate this potential bias. Third, participants willing to take part in this type of interview may be more engaged and activated than the typical patient with diabetes. Finally, we did not ask participants to specify whether they had Type 1 or Type 2 diabetes. In our clinical experience, many patients are not sure of their diabetes type and, in the absence of blood tests (i.e., insulin autoantibodies, c-peptide), providers cannot always be entirely certain of diabetes type among patients treated only with insulin. However, the age of participants and use of oral medications or diet for management suggest most participants would be classified as having type 2 diabetes.