Abstract
Objectives:
Examine the association of patient perceptions of care with hemoglobin A1c (HbA1c), medication adherence, and missed appointments in non-Hispanic Black (NHB) and White (NHW) patients with type 2 diabetes (T2DM).
Methods:
We used linear and logistic regression models to analyze the association of the Interpersonal Processes of Care survey (IPC) with HbA1c, medication adherence, and missed appointments. We then examined how these associations differed by race.
Results:
There was no overall association between IPC subdomains and HbA1c in our sample (N=221). NHB patients perceiving their provider always explained results and medications had a HbA1c on average 0.59 (−1.13, −0.04; p=0.04) points lower than those perceiving their provider sometimes explained results and medications. No effect was observed in NHWs. Never perceiving disrespect from office staff was associated with an average 0.67 (−1.1, −0.24; p=0.002) point improvement in medication adherence for all patients. Never perceiving discrimination from providers was associated with a 0.44 (−0.63, −0.25; p<0.0001) decrease in the probability of missing an appointment for NHB patients.
Conclusions:
These results demonstrate that particular aspects of communication in the patient-provider interaction may contribute to racial disparities in T2DM.
Practice Implications:
Communication training for both provider and staff may reduce disparities in T2DM.
Keywords: Type II Diabetes, Patient-Provider Interaction, Provider Communication, Medication Adherence, Glycemic Control, Healthcare disparities, Racial disparities
1. Introduction
Type 2 Diabetes Mellitus (T2DM) disproportionally affects non-Hispanic Black (NHB) populations. The estimated prevalence of diagnosed T2DM in non-Hispanic White (NHW) adults between 2013–2016 was 9.4%, however prevalence among NHB adults was significantly higher at 13.3% [1]. NHB patients also suffer worse outcomes from T2DM compared NHW patients, which are not fully explained by socioeconomic factors. Glycemic control, measured by hemoglobin A1c (HbA1c), in NHB patients remains worse than in their NHW counterparts even when controlling for access to care, insurance status and adherence to processes of care [2–4]. Additional work controlling for access to care and insurance status has found that NHB patients with diabetes are more likely to have a lower extremity amputation and develop end-stage renal disease (ESRD) than NHW patients [5–7].
The patient-provider relationship is an important and modifiable factor in the management of chronic disease. Increased patient dignity and autonomy, clearer provider communication, and improved perception of physician friendliness and respect are associated with higher reported patient adherence to physician recommendations and receipt of optimal preventive care [8–10]. In addition to improvements in health behaviors, the patient-provider relationship has also been shown to have a significant effect on health outcomes including HbA1c, low density lipoprotein (LDL), and systolic blood pressures [7,11,12].
NHB patients on average report lower trust in and have more negative communication experiences with their healthcare providers, indicating that intervening on the patient-provider relationship may improve health outcomes in this patient population [13–15]. Previous studies on whether differential ratings of the patient-provider relationship impact health outcomes in NHB patients have demonstrated promising results. For instance, examination of race-discordant, NHB patient and NHW provider, relationships found more collaborative communication was associated with greater medication adherence in hypertensive NHB patients [16].
Despite the added potential for impact, the relationship between components of the patient-provider interaction, health behaviors and outcomes, and race with T2DM has not been fully characterized. This study explores the use of the Interpersonal Processes of Care (IPC) survey, an instrument validated in diverse patient cohorts, which characterizes the patient-provider interaction along seven independent subdomains [17]. We examined the association of the IPC survey with patient race, glycemic control (HbA1c), self-reported medication adherence, and missed appointments in a population of NHB and NHW patients with T2DM.
2. Methods
2.1. Study Design
This is a cross-sectional study of patients with T2DM conducted between December 2019 and February 2020.
2.2. Cohort Identification
Eligible patients were identified using a DEDUCE1 search of the electronic medical record (EMR) within the Duke University Health System [18]. We included adult NHB and NHW patients with two or more Duke primary care visits associated with an ICD-10 code for T2DM, with the same primary care provider (PCP) in the past year. Patients also had to have at least one HbA1c measurement in the past year, be prescribed at least one daily anti-hyperglycemic medication, and could not have opted out of research. We excluded patients who had more than one appointment with multiple Duke PCPs unless there was one PCP with whom a patient had a majority of appointments (see Figure 1 for detail).
Figure 1: Patient Recruitment (color print online only).
This figure demonstrates the selection and recruitment process for our patient sample. Partially eligible patients met all eligibility criteria except prescription of ≥1 antihyperglycemic medication and were therefore included in analyses of HbA1c and missed appointments but not medication adherence. One patient completed the IPC survey but not Extent of Medication Adherence Survey (“eligible consented patients who did not complete all survey instruments”) and two patients consented but did not complete surveys (“eligible consented patients who did not complete survey”).
Given an enrollment goal of 200 patients (100 NHB and 100 NHW) and an anticipated response rate of 10%, we drew a random sample of 1983 patients from an eligible cohort of 14,847 patients.
Permission for the conduct of this study was obtained from the Institutional Review Board (IRB) and all patients underwent informed consent over the phone or online before participating.
2.3. Recruitment
We stratified our patient sample by provider to ensure variability for a separate analysis of interest and by race for the purposes of enrolling equal numbers of NHB and NHW patients. Patients were assigned a unique record ID, which was randomly distributed within strata, and were contacted in numerical order. If an email address was available, first contact attempt was made via email, followed by up to two phone calls. If no email address was available, three phone calls were made 2–5 days apart. Email contact included a brief description of the study and a link to complete the consent and survey online. If patient contact was made by phone call, the participant was provided a brief overview of the study and the option to complete the consent and survey over the phone or online. After three attempted contacts without response patients were classified as non-respondents. In order to compensate for the time spent on the surveys, all patients who enrolled were entered into a raffle to receive a $50 gift card with a 1/20 chance of winning.
2.4. Measures
Participants answered survey questions on demographic information including self-reported race and ethnicity, level of education and financial security (See Appendix A). Participants also responded to the Interpersonal Processes of Care Survey (IPC), a 29-question instrument validated in diverse cohorts to characterize patient perceptions of care along seven subdomains, each of which serves as an independent predictor [17]. There are three negative subdomains—1. “hurried communication,” 2. “discriminated,” and 3. “disrespectful office staff”—and four positive subdomains—1. “elicited concerns and responded,” 2. “explained results and medications,” 3. “patent centered decision making,” and 4. “compassionate and respectful.” The score for each subdomain is calculated by averaging the responses ranging from 1 (never) to 5 (always) along a Likert scale for the questions within that subdomain. Thus, the score for each subdomain ranges between 1–5. For negatively scored subdomains such as “hurried communication” lower scores indicate more favorable perceptions (i.e. less frequent hurried communication). For positive subdomains such as “elicited concerns and responded” higher scores indicate a more favorable perception (i.e. greater frequency of eliciting concerns).
Information on comorbidities as well as type (oral versus subcutaneous injection) and number of anti-hyperglycemic medications was collected from the EMR. Diabetes-related neuropathy, retinopathy, and nephropathy (classified as CKD Stage 3 and above) were assessed through chart review of ICD-10 codes in the patient problem list. Patients were categorized as having experienced a heart attack or stroke if they self-reported either outcome, or if myocardial infarction (MI), cerebrovascular accident (CVA), or transient ischemic attack (TIA) were found in their problem list, except for two participants with missing self-report data for whom only chart review was used.
2.4.1. Outcome Measures
Our outcomes of interest were HbA1c, medication adherence, and missed appointments. Most recent HbA1c was recorded from the EMR. Medication adherence was elicited via patient self-report through the Extent of Adherence Survey, a 3-item tool with scores ranging from 1–5 for which lower scores indicate greater adherence, that is fewer missed doses [19]. Missed appointment information was obtained from Business Objects Web Intelligence2 and was calculated by combining cancelled and not rescheduled appointments and no-showed appointments. Missed appointments were then dichotomized at no missed appointments or at least one missed appointment
2.5. Data Storage
REDCap, a secure database with survey capability, was used to build out our survey instrument and to collect and store participant data [20,21].
2.6. Statistical Analyses
Demographic variables were tabulated overall and by subgroup. Gender, age, race, most recent HbA1c, and missed appointments were tabulated for non-respondents.
Due to lack of variability in response, scores for each subdomain of the IPC Survey were dichotomized at fully endorsed and did not fully endorse. We defined fully endorsed as patients responding to all answers within a subdomain with the most favorable response. For positive domains like elicited concerns, fully endorsed refers an answer of “always” to all items within the domain. In describing the results, we refer to these responses as, for example, “always elicited concerns.” For negative domains, such as disrespectful office staff, fully endorsed refers to an answer of “never” to all items within the domain. In describing the results, we refer to these responses as, for example, “never disrespectful office staff.” If not all items were fully endorsed within a subdomain, we refer to the responses as “sometimes,” that is “sometimes discriminated” or “sometimes patient centered decision making.”
We conducted two sets of multiple regression analyses to examine the association between IPC domains and our outcomes of interest. The first set of models included each IPC subdomain and self-reported race and were adjusted for age, gender, length of relationship with provider, financial security and education. Medication adherence models were also adjusted for route of medication administration. The second set of models additionally included the interaction between IPC subdomain and race, as one of our primary research questions was to understand the potential differential impact of IPC for NHB patients as compared to NHW patients on outcomes. Multivariable linear regression was used for HbA1c and medication adherence, and we a priori defined clinically significant differences as a 0.5% mean change in HbA1c, as is standard for clinical trials [22], and a 1-point mean change on the Extent of Adherence Survey which indicates an improvement of a full category of adherence on the scale. Multivariable logistic regression was used to examine the association of IPC response and missed appointments, and for ease of interpretation, results are reported as estimated probabilities of experiencing at least one missed appointment. We determined that a 25 percentage point reduction in the probability of missing an appointment was a meaningful clinical cut off given that patients with T2DM are typically have between 2–4 PCP visits per year, thus a 25% reduction would equate to one less missed visit per year for patients requiring the closest monitoring.
SAS 9.4 (Cary, NC) was used for all analyses, and statistical significance was defined as an alpha level of 0.05. Adjustment for multiple comparisons was not performed as all models were specified a priori and all analyses and p-values are reported.
3. Results
3.1. Sample Characteristics
Our final sample consisted of 221 patients—106 NHW and 115 NHB. Twenty patients were excluded from analyses of medication adherence—nineteen patients (13 NHB and 6 NHW) who were found on chart review to no longer be prescribed anti-hyperglycemic medication at the time of survey and one NHB patient who did not complete the Extent of Adherence Survey. Thus, the analysis of medication adherence was performed on a subset of 201 patients (100 NHW and 101 NHB) out of 221 total patients (see Figure 1).
Overall our sample was majority female (55.2%) with a mean age of 64 (SD 11.2). A greater number of NHW patients were male, had a post high-school education, were insured through Medicare, reported a shorter relationship (<3 years) with their current provider, and were prescribed three or more antihyperglycemic medications (Table 1). NHB and NHW patients had similar comorbidities although more NHW patients had a history of MI.
Table 1:
Characteristics of the patient population
| Overall N=221 | White N=106 | Black N=115 | |
|---|---|---|---|
| Age, mean (SD) | 64.4 (11.2) | 65.4 (11.4) | 63.5 (10.9) |
| Male, n (%) | 99 (44.8) | 64 (60.4) | 35 (30.4) |
| Insurance, n (%) | |||
| Medicaid | 5 (2.3) | 0 (0) | 5 (4.3) |
| Medicare | 75 (33.9) | 45 (42.4) | 30 (26.1) |
| Private | 83 (37.6) | 38 (35.8) | 45 (39.1) |
| Mixed/Both Private and Public | 58 (26.2) | 23 (21.7) | 35 (30.4) |
| Financial Securitya,b, n (%) | |||
| High | 85 (38.5) | 49 (46.2) | 36 (31.3) |
| Medium | 72 (32.6) | 28 (26.4) | 44 (38.3) |
| Low | 62 (28.1) | 27 (25.5) | 35 (30.4) |
| Education, n (%) | |||
| High school or less | 53 (24) | 16 (15.1) | 37 (32.2) |
| Some post-secondary | 92 (41.6) | 47 (44.3) | 45 (39.1) |
| Bachelor’s degree or greater | 76 (34.4) | 43 (40.6) | 33 (28.7) |
| Length of provider relationship, n (%) | |||
| <1 yr | 19 (8.6) | 9 (8.5) | 10 (8.7) |
| 1–3yr | 83 (37.6) | 45 (42.5) | 38 (33) |
| > 3 yr | 119 (53.8) | 52 (49.1) | 67 (58.3) |
| Presence of comorbidities, n (%) | |||
| Neuropathy | 44 (19.9) | 21 (19.8) | 23 (20) |
| Chronic Kidney Disease | 22 (10.0) | 12 (11.3) | 10 (8.7) |
| Retinopathy | 13 (5.9) | 6 (5.7) | 7 (6.1) |
| Previous MI | 24 (10.9) | 15 (14.2) | 9 (7.8) |
| Previous CVA or TIA | 27 (12.2) | 12 (11.3) | 15 (13.0) |
| Number of antihyperglycemic medicationsc, n (%) | |||
| 1 | 84 (41.6) | 38 (38) | 46 (45.1) |
| 2 | 69 (34.2) | 32 (32) | 37 (36.3) |
| 3+ | 49 (24.3) | 30 (30) | 19 (18.6) |
| Route of medication administrationc, n (%) | |||
| Oral | 121 (59.9) | 63 (63) | 58 (56.9) |
| Subcutaneous | 17 (8.4) | 6 (6) | 11 (10.8) |
| Both | 64 (31.7) | 31 (31) | 33 (32.4) |
| Number of appointments with provider in the past year, median (Q1, Q3) | 5 (4, 8) | 5 (4, 8) | 5 (4, 7) |
N=104 White, 115 Black due to non-response
Financial security was categorized as high, medium or low using the following prompts: after paying the bills you still have enough money for special things that you want (high); you have enough money to pay the bills, but little spare money to buy extra or special things (medium); you have money to pay the bills, but only because you have cut back on things (low); and you are having difficulty paying the bills, no matter what you do (low).
N= 100 White, 102 Black
NHB and NHW patients answered similarly on the IPC subdomains except for the subdomains “Disrespectful Office Staff” and “Discriminated” which NHW patients reported experiencing at a higher frequency (Table 2). We did not find a significant difference in glycemic control (HbA1c) between groups, on average 7.54 (SD 1.67) in NHB patients and 7.32 (SD 1.25) in NHW patients (p=0.3), or medication adherence, average score 2.2 (SD 1.06) in NHB and 2.02 (SD 1.03) in NHW patients (p=0.23). A higher percentage of NHB patients had missed at least one appointment in the past year, but this difference was not significant (p=0.11).
Table 2: Outcome measures in the patient population.
This table displays the number and percentage of patients endorsing the most favorable response on the IPC Survey and the results for our outcomes of interest.
| Measure | Overall | Non-Hispanic White | Non-Hispanic Black | p-value | |
|---|---|---|---|---|---|
| IPC Subdomain, n (%) | Never Hurried Communication | 119 (53.4) | 56 (52.8) | 63 (54.8) | 0.77 |
| Never Discriminated | 174 (78.7) | 75 (70.7) | 99 (86.1) | 0.005 | |
| Never Disrespectful Office staff | 185 (83.7) | 80 (75.5) | 105 (91.3) | 0.002 | |
| Always Elicited Concerns and Responded | 139 (62.9) | 64 (60.4) | 75 (65.2) | 0.46 | |
| Always Explained Results and Medications | 89 (40.3) | 37 (34.9) | 52 (45.2) | 0.12 | |
| Always Patient Centered Decision Makinga | 74 (33.5) | 38 (36.2) | 36 (31.3) | 0.44 | |
| Always Compassionate, respectful | 136 (61.5) | 60 (56.6) | 76 (66.1) | 0.15 | |
| Most Recent HbA1c, mean (SD) | 7.32 (1.48) | 7.32 (1.25) | 7.54 (1.67) | 0.3 | |
| Average medication adherence scoreb, mean (SD) | 2.11 (1.04) | 2.02 (1.03) | 2.20 (1.06) | 0.23 | |
| ≥1 missed appointment in the past year, n (%) | 100 (45.2) | 42 (39.6) | 58 (50.4) | 0.11 | |
N=220 due to nonresponse
N=201
We attempted at least one contact with 1244 patients. 675 were classified as non-respondents, three attempted contacts with no response or opted out on contact, for a response rate of 23.6% and 26% for NHB and NHW patients respectively. Nonrespondents were similar in age, gender, HbA1c and missed appointments to our sample—mean age 65 (SD 13.3), 55% female, mean recent HbA1c of 7.4, and a 46.1% had missed an appointment in the past year.
3.3. Hemoglobin A1c
We did not observe any clinically or statistically significant association of IPC subdomain with HbA1c in the adjusted regression models without the race interaction term (Appendix B Table 1). In the set of models including the interaction of IPC subdomain and race, we observed a clinically significant interaction in the subdomain of explained results and medications and hurried communication. NHB patients reporting that their provider always explained results and medications had a mean estimated HbA1c that was 0.59 (−1.13, −0.04; p=0.04) percentage points lower than NHB respondents reporting their provider sometimes explained results and medications. This effect was the opposite in NHW patients who had a mean estimated HbA1c that was 0.4 (−0.21, 1.00; p=0.2) percentage points higher when reporting that their provider always explained results and medications as compared to sometimes explaining them. Thus, for NHB patients the differential effect of reporting a provider always explained results and medications was a mean estimated HbA1c 0.98 (−1.80, −0.16; p=0.02) percentage points lower than NHW patients. Additionally, NHB patients who reported never experiencing hurried communication with their provider had an estimated mean HbA1c 0.46 (−1.0, 0.09; p=0.10) percentage points lower than NHB patients who sometimes experienced hurried communication from their provider. In contrast, NHW patients who reported never experiencing hurried communication had an estimated mean HbA1c 0.36 (−0.22, 0.93; p=0.23) percentage points higher than NHW patients reporting sometimes experiencing hurried communication. The mean estimated differential effect for NHB patients of never experiencing hurried communication was a HbA1c 0.82 (−1.62, −0.01, p=0.05) percentage points lower than NHW patients (Figure 2). Overall, these findings indicate that for NHB patients always having results and medications explained and never experiencing hurried communication from a provider were associated with substantial increases in glycemic control, but this relationship was not observed in NHWs.
Figure 2:
Relationship between specified IPC subdomains and HbA1c by race
* Adjusted for age, gender, race, length of relationship with provider, education, and financial security
3.4. Medication Adherence
In the set of adjusted regression models for medication adherence without the race interaction term, we observed that the subdomain of disrespectful office staff was associated with lower scores on the Extent of Adherence Survey. Lower scores indicate better adherence to taking medications as prescribed. In the overall sample, patients who reported that office staff were never disrespectful had an estimated mean score 0.67 (−1.1, −0.24; p=0.002) points lower than those reporting office staff were sometimes disrespectful. We did not observe any significant associations in the set of models which included the interaction of IPC subdomain and race (See Appendix B Table 2). However, we did observe that the association between the perception of disrespectful office staff and medication adherence was more pronounced in NHB patients. For NHB patients reporting never disrespectful office staff, estimated mean score on the Extent of Adherence Survey was 1.04 (−1.8, −0.28; p=0.01) points lower, meeting our threshold for clinically meaningful change, compared to NHW patients with an estimated mean score 0.52 (−1.03, −0.01; p=0.04) points lower (Figure 3). Taken as a whole, patients who perceived disrespect from office staff had worse medication adherence than those that never perceived disrespect from office staff.
Figure 3:
Relationship between disrespectful office staff and medication adherence by race
* Adjusted for age, gender, race, length of relationship with provider, education, financial security, and route of antihyperglycemic medication administration.
3.5. Missed Appointments
We observed in the set of adjusted models without the race interaction term that never discriminating was associated with a clinically significant, greater than 0.25 difference, 0.28 decrease (−0.44, −0.11; p=0.001) in the estimated probability of missing an appointment. In the set of adjusted models with the race interaction term we found a statically and clinically significant association between missing an appointment and the subdomain “discriminated” and a clinically significant difference for the subdomain “disrespectful office staff” (See Appendix B Table 3 for all probabilities). In the subdomain “discriminated” the estimated probability of missing an appointment in a year was 0.42 for NHB patients reporting never experiencing discrimination from their provider and was 0.86 for NHB patients reporting sometimes experiencing discrimination (difference= −0.44, 95% CI: −0.63, −0.25; p<0.0001). Among NHW patients, this difference was −0.15 (−0.37, 0.08; p=0.20). The differential effect of never discriminating for NHB compared to NHW patients upon the probability of a missed appointment was −0.30 (−0.58, −0.007; p=0.04) (See Fig. 4).
Figure 4:
Predicted probability of missing an appointment for specified IPC subdomains*
*Adjusted for age, gender, race, length of relationship with provider and education
Additionally, among NHW patients, the estimated probability of missing an appointment in a year was 0.35 for those who reported never disrespectful office staff compared to 0.58 for those who reported sometimes experiencing disrespectful office staff (difference= −0.23, 95% CI: −0.46, 0; p=0.05). In contrast, among NHB patients, this difference was 0.14 (−0.18, 0.46; p=0.39). The differential effect of disrespectful office staff on NHW patients was a 0.37 (−0.01, 0.76; p=0.06) lower probability of missing an appointment when never experiencing disrespectful office staff. Overall, never experiencing discrimination was associated with a lower probability of missing an appointment in both NHW and NHB patients and never experiencing disrespect from office staff was associated with a lower probability of missing an appointment in NHW patients.
4. Discussion and Conclusion
4.1. Discussion
This study examines the interplay of patient perceptions of care, race, and health outcomes in NHB and NHW patients with T2DM. In this study we found two important interactions between race and IPC survey response on HbA1c. We observed a mean estimated −0.98 and −0.81 percentage point differential effect for NHB patients compared to NHW patients who reported results and medications were always explained and their provider never hurried communication respectively. These findings are in concordance with previous work showing that patient provider interactions can play an important role in improving management of and outcomes from diabetes including glycemic control [2,11,12,23].
However, we were surprised to observe that for NHW patients, more positive ratings of providers in these subdomains on the IPC were associated with higher HbA1c. We hypothesize that this finding may be due in part to other factors in the patient provider interaction that make optimal communication and explanation less important to NHW patients in achieving glycemic control. Thirty percent of NHW patients in our sample were prescribed three or more antihyperglycemic medications compared to 19% of NHB patients. This indicates that factors such as increased therapeutic intensification for NHW patients may underly the lack of association seen between IPC response and HbA1c in NHW patients [24]. Additionally, previous work has shown that NHB patients are less likely to engage in information eliciting behaviors especially in race-discordant provider interactions [25]. This difference in communication style between NHB and NHW patients may also increase the importance of provider-initiated, unhurried, and clear explanation and information provision for NHB patients in achieving optimal glycemic control.
We had not predicted that the subdomain “disrespectful office staff” would emerge as the only statistically and clinically significant association between IPC subdomain and medication adherence. Previous work characterizing the association between the IPC and adherence to insulin injections found that the subdomains of discrimination, hurried communication, and explained results and medication were associated with adherence behavior, but did not examine the subdomain of disrespectful office staff [23]. In qualitative work by Cuevas et al., Black patients with chronic disease describe perceived disrespect from office staff as another form of discrimination in the encounter [26,27]. Thus, while previous work found an association with the subdomain “discriminated,” in our study it may be that discriminatory/disrespectful experiences were reported under the subdomain “disrespectful office staff” in our study when this option was made available. Additionally, perceived unfairness of treatment, which overlaps with both the subdomains of disrespectful office staff and provider discrimination, has been correlated with delays in medication pick-up from pharmacies in a diverse patient population, providing a theoretical model for how this affects medication adherence [28]. As T2DM is a chronic disease that requires consistent office visits and interactions with staff at and between visits, it may be that office staff play an especially important role in shaping patient perceptions of their diabetes care plans and comfort with treatment. In our study the effect of never versus sometimes perceiving disrespect from office staff was particularly large for NHB patients where a full point difference on the Extent of Adherence Survey was observed, indicating that this may be a pertinent and understudied point of intervention to improve medication adherence.
We found that NHW patients who reported never perceiving disrespect were less likely to miss an appointment than those who sometimes perceived disrespect from office staff. However, this same trend was not observed in NHB patients. In his interviews, Cuevas documents a theme of NHB patients becoming habituated to disrespect in the medical encounter [27]. We therefore interpret this finding with caution as this same association may not be present in NHB patients due to under-reporting in this group, and interventions targeted at decreasing perceived disrespect from office staff may still improve experience and appointment attendance for both NHW and NHB patients. We did observe a strong association between the subdomain of discrimination and missing an appointment including a striking 0.44 difference in average predicted probability of missing an appointment between NHB patients who reported that their provider never discriminated and those who reported their provider sometimes discriminated. Previous work in a similar patient population in Durham, NC has also shown that perceived racism in the health care system is associated with patient delay of treatment [28]. This difference in probability of missing an appointment could significantly affect continuity of care and ability to achieve glycemic control.
4.2. Limitations
This study has a number of limitations. It was an exploratory study with a relatively small sample size of 221 patients. However, it was strengthened by stratifying recruitment to include equal numbers of NHW and NHB participants. Our sample had some surprising characteristics. We found low variability in IPC response which led us to dichotomize at fully endorse and did not fully endorse. Although we did not investigate reasons for low variability, we hypothesize that this may be because patients who opt-in to research have a better overall relationship with the healthcare system and their provider. Further, we did not find a significant difference in HbA1c between NHB and NHW patients in our cohort contrary to previously documented disparities [29]. One possible explanation for this discrepancy is that our sample consisted of patients with recent, within the past year, PCP visits who were networked into care. Comorbidities were also distributed similarly between the two groups except for incidence of MI, which was higher among NHW patients. This also differs from existing work demonstrating that Black Americans bear a higher burden of sequalae from T2DM except for cardiac events where there are similar rates between NHB and NHW patients [6]. We hypothesize that higher rate of cardiac events in NHW patients in our sample may be partially due to known under-diagnosis of coronary events in both NHB patients and women, who were over-represented in our NHB sample [30,31].
Notably, the only significant differences in our sample between NHB and NHW patients in IPC response were in the subdomains disrespectful office staff and discriminated which NHW patients reported experiencing more frequently. As discussed above this may be in part due to under-reporting by NHB patients or greater sensitivity to perceived discrimination and disrespect among NHW patients. In line with our findings, similar, or even more positive, responses by Black patients compared to NHW patients on the IPC Survey have been previously observed [32,33]. However, recordings of clinic visits have demonstrated objective differences in the patient-provider interaction including decreased patient-centeredness, increased physician dominance and decreased positive emotional affect for Black patients [15]. Additionally, there is a large body of work documenting decreased trust in the patient-provider relationship for Black patients [13,14,34,35]. Our findings may be indicative of the fact that our cohort is comprised of patients who self-selected to participate in this research and have greater access to care and possibly stronger relationships with their providers than the average population. A broader community-based sample may have yielded a wider range of responses and comfort sharing negative aspects of the patient-provider relationship such as in previous broad large telephone surveys [36].
4.3. Conclusions
In this study we found that slower pace of communication and greater clarity of explanation were associated with achieving better glycemic control in NHB patients. We further observed that aspects of the clinic environment such as perceived disrespectfulness of office staff and discrimination from one’s provider were associated with worse medication adherence and higher probability of missing an appointment. Despite limitations in the research design, these findings highlight specific areas of the patient-provider relationship for future intervention to impact these important clinical outcomes, particularly for NHB patients.
4.4. Practice Implications
This study suggests that there are multiple aspects of the patient-provider and patient-staff interaction that may serve as promising intervention points to improve disparities in care for NHB patients with T2DM. Practices and health systems should focus on training opportunities to improve not only patient-provider communication but also patient-office staff communication and interaction. Training to improve patient-provider communication and patient-office staff communication and interaction may lead to improved patient-centered care and outcomes in T2DM especially for NHB patients.
Highlights.
Explanation of results and medications is associated with better glycemic control.
Hurried communication is associated with worse glycemic control.
The above associations are present for non-Hispanic Black patients, not Whites.
Disrespectful office staff are associated with poor diabetes medication adherence.
Discrimination is associated with more missed appointments.
Acknowledgements:
Thank you to Shawin Vitsupakorn for his help with patient recruitment. Research reported in this publication was supported by the Duke Center for Research to Advance Health Equity under Award Number 5U54MD012530–03 and by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number TL1 TR002555.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the Duke Center for Research to Advance Health Equity or the National Institutes of Health.
Appendix A: Survey Instruments
A.1.
Demographic Questionnaire
| Demographic Questionnaire | |
|---|---|
| Today’s Date | |
| Please select the race you most closely identify with. Please note that only Black and White patients are eligible to participate in this study |
|
| Please select the ethnicity you most closely identify with. Please note that only non-Hispanic/Latino patients are eligible to participate in this study |
|
| How long have you been receiving care from your current primary care provider? By this we mean the physician, nurse practitioner, or physician’s assistant who cares for your diabetes (high blood sugar). |
|
| What is the highest level of education you have completed? |
|
| Without giving exact dollars how would you describe your household’s financial situation right now? Would you say: |
|
| Have you ever had a heart attack? |
|
| Have you ever had a stroke? |
|
Appendix B: Additional Analyses
Table 1: Difference in HbA1c by IPC Subdomain Response.
This table displays estimated difference HbA1c adjusted for age, gender, race, length of relationship with provider, education, and financial security for respondents answering either always/never versus sometimes for a given IPC subdomain. Race is coded with NHW as the referent category.
| IPC Sub-Domain | Difference Between Never and Sometimes1 | NHW Difference Between Never and Sometimes2 | NHB Difference Between Never and Sometimes2 | Difference NHB versus NHW2 |
|---|---|---|---|---|
| Hurried communication | −0.07 (−0.47, 0.33) p=0.74 | 0.36 (−0.22, 0.93) p=0.23 | −0.46 (−1.02, 0.09) p=0.10 | −0.82 (−1.62, − 0.01) p=0.05 |
| Discriminated | 0.01 (−0.50, 0.52) p=0.97 | −0.001 (−0.65, 0.65) p=1 | 0.03 (−0.79, 0.84) p=0.95 | 0.03 (−1.02, 1.07) p=0.96 |
| Disrespectful office staff | 0.35 (−0.22, 0.91) p=0.23 | 0.4 (−0.28, 1.08) p=0.25 | 0.22 (−0.77, 1.22) p=0.66 | −0.18 (−1.36, 1.01) p=0.77 |
| IPC Sub-Domain | Difference Between Always and Sometimes1 | NHW Difference Between Always and Sometimes | NHB Difference Between Always and Sometimes | Difference NHB versus NHW |
| Elicited concerns, responded | −0.03 (−0.45, 0.39) p=0.88 | 0.10 (−0.50, 0.70) p=0.74 | −0.16 (−0.74, 0.42) p=0.59 | −0.26 (−1.09, 0.57) p=0.54 |
| Explained results, medications | −0.15 (−0.56, 0.27) p=0.49 | 0.40 (−0.21, 1.0) p=0.20 | −0.59 (−1.14, −0.04) p=0.04 | −0.98 (−1.80, −0.16) p=0.02 |
| Patient-centered decision making | 0.07 (−0.36, 0.49) p=0.76 | 0.22 (−0.39, 0.83) p=0.48 | −0.08 (−0.68, 0.52) p=0.79 | −0.30 (−1.16, 0.56) p=0.49 |
| Compassionate, respectful | 0.21 (−0.21, 0.62) p=0.33 | 0.36 (−0.23, 0.95) p=0.23 | 0.05 (−0.53, 0.63) p=0.86 | −0.31 (−1.14, 0.52) p=0.46 |
These results are from the first set of analyses which include all covariates without the interaction term
These results are from the second set of analyses which include all covariates as well as the IPC × race interaction term.
Table 2: Difference in Medication Adherence by IPC Subdomain Response.
This table displays estimated difference in medication adherence as measured by the Extend of Adherence Survey adjusted for age, gender, race, length of relationship with provider, education, financial security, and route of medication administration for respondents answering either always/never versus sometimes for a given IPC subdomain. Race is coded with NHW as the referent category.
| IPC Sub-Domain | Difference Between Never and Sometimes1 | NHW Difference Between Never and Sometimes2 | NHB Difference Between Never and Sometimes2 | Difference NHB versus NHW2 |
|---|---|---|---|---|
| Hurried communication | −0.17 (−0.47, 0.12) p=0.25 | −0.32 (−0.75, 0.10) p=0.14 | −0.03 (−0.45, 0.40) p=0.90 | 0.30 (−0.31, 0.90) p=0.33 |
| Discriminated | −0.09 (−0.47, 0.29) p=0.65 | −0.07 (−0.55, 0.42) p=0.79 | −0.12 (−0.74, 0.50) p=0.70 | −0.05 (−0.85, 0.74) p=0.89 |
| Disrespectful office staff | −0.67 (−1.1, −0.24) p=0.002 | −0.52 (−1.03, −0.01) p=0.04 | −1.04 (−1.8, −0.27) p=0.01 | −0.51 (−1.42, 0.39) p=0.26 |
| Difference Between Always and Sometimes1 | NHW Difference Between Always and Sometimes2 | NHB Difference Between Always and Sometimes2 | Difference NHB versus NHW2 | |
| Elicited concerns, responded | −0.17 (−0.48, 0.14) p=0.29 | −0.16 (−0.59, 0.28) p=0.47 | −0.18 (−0.62, 0.26) p=0.43 | −0.02 (−0.64, 0.60) p=0.95 |
| Explained results, medications | −0.17 (−0.47, 0.14) p=0.29 | −0.01 (−0.46, 0.44) p=0.96 | −0.30 (−0.72, 0.12) p=0.16 | −0.29 (−0.90, 0.33) p=0.36 |
| Patient-centered decision making | −0.26 (−0.58, 0.06) p=0.11 | −0.16 (−0.61, 0.28) p=0.48 | −0.36 (−0.82, 0.09) p=0.11 | −0.20 (−0.84, 0.43) p=0.53 |
| Compassionate, respectful | −0.15 (−0.46, 0.16) p=0.34 | −0.18 (−0.62, 0.25) p=0.4 | −0.11 (−0.56, 0.33) p=0.62 | 0.07 (−0.55, 0.69) p=0.82 |
These results are from the first set of analyses which include all covariates without the interaction term
These results are from the second set of analyses which include all covariates as well as the IPC × race interaction term.
Table 3: Difference in Probability of Missing an Appointment by IPC Subdomain Response.
This table shows the difference in predicted probability of missing an appointment adjusted for age, gender, race, length of relationship with provider, education, and financial security for respondents answering never/always versus sometimes in a given IPC subdomain.
| IPC Sub-Domain | Difference Between Never and Sometimes1 | NHW Difference Between Never and Sometimes2 | NHB Difference Between Never and Sometimes2 | Difference NHB versus NHW2 |
|---|---|---|---|---|
| Hurried communication | 0.05 (−0.09, 0.19) p=0.45 | 0.02 (−0.18, 0.21) p=0.86 | 0.08 (−0.11, 0.28) p=0.39 | 0.07 (−0.21, 0.34) p=0.64 |
| Discriminated | −0.28 (−0.44, −0.11) p=0.001 | −0.15 (−0.37, 0.08) p=0.20 | −0.44 (−0.63, − 0.25) p<0.0001 | −0.30 (−0.58, − 0.007) p=0.04 |
| Disrespectful office staff | −0.11 (−0.31, 0.08) p=0.25 | −0.23 (−0.46, − 0.003) p=0.047 | 0.14 (−0.18, 0.46) p=0.39 | 0.37 (−0.02, 0.76) p=0.06 |
| IPC Sub-Domain | Difference Between Always and Sometimes1 | NHW Difference Between Always and Sometimes2 | NHB Difference Between Always and Sometimes2 | Difference NHB versus NHW2 |
| Elicited concerns, responded | 0.012 (−0.13, 0.15) p=0.86 | −0.02 (−0.22, 0.18) p=0.83 | 0.05 (−0.15, 0.24) p=0.66 | 0.07 (−0.21, 0.35) p=0.64 |
| Explained results, medications | 0.02 (−0.13, 0.16) p=0.83 | 0.11 (−0.10, 0.31) p=0.32 | −0.06 (−0.25, 0.13) p=0.55 | −0.16 (−0.45, 0.12) p=0.25 |
| Patient-centered decision making | 0.11 (−0.04, 0.25) p=0.15 | 0.09 (−0.12, 0.30) p=0.39 | 0.12 (−0.08, 0.33) p=0.23 | 0.03 (−0.26, 0.32) p=0.82 |
| Compassionate, respectful | 0.002 (−0.14, 0.14) p=0.98 | 0.10 (−0.10, 0.29) p=0.34 | −0.09 (−0.29, 0.11) p=0.38 | −0.18 (−0.46, 0.010) p=0.20 |
These results are from the first set of analyses which include all covariates without the interaction term
These results are from the second set of analyses which include all covariates as well as the IPC × race interaction term.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of interest/Competing interests: None
Ethics approval: The questionnaires and methodology for this study were approved by the Institutional Review Board (IRB) at Duke University Hospital.
Availability of data and material: De-identified data can be made available upon request with the appropriate data-sharing agreements in place.
Code availability: Code can be made available upon request with the appropriate data-sharing agreements in place.
The Duke Enterprise Data Unified Content Explorer (DEDUCE) is a web-based tool that allows the creation of study cohorts based on criteria available in clinical records.
Business Objects Web Intelligence is a data warehouse containing appointment and financial metrics concerning Duke affiliated providers and clinics.
References
- [1].CDC, National Diabetes Statistics Report 2020. Estimates of diabetes and its burden in the United States., 2020.
- [2].Narayan KMV, Duru OK, Selby JV, Swain BE, Ferrara A, Waitzfelder BE, Ackermann RT, Gerzoff RB, Caputo D, Brown AF, Herman W, Marrero DG, Karter AJ, Understanding the Gap Between Good Processes of Diabetes Care and Poor Intermediate Outcomes, Med. Care 45 (2009) 1144–1153. 10.1097/mlr.0b013e3181468e79. [DOI] [PubMed] [Google Scholar]
- [3].Harris MI, Eastman RC, Cowie CC, Flegal KM, Eberhardt MS, Racial and ethnic differences in glycemic control of adults with type 2 diabetes, Diabetes Care. 22 (1999) 403–408. 10.2337/diacare.22.3.403. [DOI] [PubMed] [Google Scholar]
- [4].Bulger JB, Shubrook JH, Snow R, Racial disparities in African Americans with diabetes: Process and outcome mismatch, Am. J. Manag. Care 18 (2012) 407–413. [PubMed] [Google Scholar]
- [5].Young BA, Maynard C, Reiber G, Boyko EJ, Effects of ethnicity and nephropathy on lower-extremity amputation risk among diabetic veterans, Diabetes Care. 26 (2003) 495–501. 10.2337/diacare.26.2.495. [DOI] [PubMed] [Google Scholar]
- [6].Golden SH, Brown A, Cauley JA, Chin MH, Gary-Webb TL, Kim C, Sosa JA, Sumner AE, Anton B, Health disparities in endocrine disorders: Biological, clinical, and nonclinical factors - An endocrine society scientific statement, J. Clin. Endocrinol. Metab 97 (2012) E1579. 10.1210/jc.2012-2043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Karter AJ, Ferrara A, Liu JY, Moffet HH, Ackerson LM, Selby JV, Ethnic disparities in diabetic complications in an insured population, J. Am. Med. Assoc 287 (2002) 2519–2527. 10.1001/jama.287.19.2519. [DOI] [PubMed] [Google Scholar]
- [8].Dimatteo MR, Haskard Zolnierek KB, Physician Communication and Patient Adherence to Treatment: A Meta-analysis, Med. Care 47 (2009) 826–834. 10.1097/MLR.0b013e31819a5acc.Physician. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Beach MC, Sugarman J, Johnson RL, Arbelaez JJ, Duggan PS, Cooper LA, Do patients treated with dignity report higher satisfaction, adherence, and receipt of preventive care?, Ann. Fam. Med 3 (2005) 331–338. 10.1370/afm.328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Sherbourne CD, Hays RD, Ordway L, DiMatteo MR, Kravitz RL, Antecedents of adherence to medical recommendations: Results from the medical outcomes study, J. Behav. Med 15 (1992) 447–468. 10.1007/BF00844941. [DOI] [PubMed] [Google Scholar]
- [11].Lafata JE, Morris HL, Doble E, Heisler Mi., Werner RM, Dumenci L, Patient-reported use of collaborative goal setting and glycemic control among patients with diabetes, Patient Educ Couns. 92 (2013) 94–99. 10.1016/j.pec.2013.01.016.Patient-reported. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Mancuso JM, Impact of health literacy and patient trust on glycemic control in an urban USA population, Nurs. Heal. Sci 12 (2010) 94–104. 10.1111/j.1442-2018.2009.00506.x. [DOI] [PubMed] [Google Scholar]
- [13].Musa D, Schulz R, Harris R, Silverman M, Thomas SB, Trust in the health care system and the use of preventive health services by older black and white adults, Am. J. Public Health 99 (2009) 1293–1299. 10.2105/AJPH.2007.123927. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Halbert CH, Armstrong K, Gandy OH Jr, Shaker L, Racial Differences in Trust in Health Care Providers, Arch. Intern. Med 166 (2006) 896–901. http://archinte.ama-assn.org/cgi/content/abstract/166/8/896. [DOI] [PubMed] [Google Scholar]
- [15].Johnson RL, Roter D, Powe NR, Cooper LA, Patient Race/Ethnicity and Quality of Patient–Physician Communication During Medical Visits, Am. J. Public Health 94 (2004) 2084–2090. 10.2105/AJPH.94.12.2084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Schoenthaler A, Allegrante JP, Chaplin W, Ogedegbe G, The effect of patient-provider communication on medication adherence in hypertensive black patients: Does race concordance matter?, Ann. Behav. Med 43 (2012) 372–382. 10.1007/s12160-011-9342-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Stewart AL, Nápoles-Springer AM, Gregorich SE, Santoyo-Olsson J, Interpersonal processes of care survey: Patient-reported measures for diverse groups, Health Serv. Res 42 (2007) 1235–1256. 10.1111/j.1475-6773.2006.00637.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Horvath MM, Winfield S, Evans S, Slopek S, Shang H, Ferranti J, The DEDUCE Guided Query tool: Providing simplified access to clinical data for research and quality improvement, J. Biomed. Inform 44 (2011) 266–276. 10.1016/j.jbi.2010.11.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Voils CI, MacIejewski ML, Hoyle RH, Reeve BB, Gallagher P, Bryson CL, Yancy WS, Initial validation of a self-report measure of the extent of and reasons for medication nonadherence, Med. Care 50 (2012) 1013–1019. 10.1097/MLR.0b013e318269e121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG, Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support, J. Biomed. Inform 42 (2009) 377–381. 10.1016/J.JBI.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, McLeod L, Delacqua G, Delacqua F, Kirby J, Duda SN, The REDCap consortium: Building an international community of software platform partners, J. Biomed. Inform 95 (2019) 103208. 10.1016/j.jbi.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Little RR, Rohlfing CL, The long and winding road to optimal HbA1c measurement, Clin. Chim. Acta 418 (2013) 63–71. 10.1016/j.cca.2012.12.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Linetzky B, Jiang D, Funnell MM, Curtis BH, Polonsky WH, Exploring the role of the patient-physician relationship on insulin adherence and clinical outcomes in type 2 diabetes: Insights from the MOSAIc study, J. Diabetes 9 (2017) 596–605. 10.1111/1753-0407.12443. [DOI] [PubMed] [Google Scholar]
- [24].Rodondi N, Peng T, Karter AJ, Bauer DC, Vittinghoff E, Tang S, Pettitt D, Kerr EA, Selby JV, Therapy modifications in response to poorly controlled hypertension, dyslipidemia, and diabetes mellitus, Ann. Intern. Med 144 (2006) 475–484. 10.7326/0003-4819-144-7-200604040-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Gordon HS, Street RL, Sharf BF, Souchek J, Racial differences in doctors’ information-giving and patients’ participation, Cancer. 107 (2006) 1313–1320. 10.1002/cncr.22122. [DOI] [PubMed] [Google Scholar]
- [26].Cuevas AG, O’Brien K, Saha S, African American experiences in healthcare: “I always feel like I’m getting skipped over,” Heal. Psychol 35 (2016) 987–995. 10.1037/hea0000368. [DOI] [PubMed] [Google Scholar]
- [27].Cuevas AG, Exploring Four Barriers Experienced by African Americans in Healthcare: Perceived Discrimination, Medical Mistrust, Race Discordance, and Poor Communication, Portland State University, 2013. [Google Scholar]
- [28].Van Houtven CH, Voils CI, Oddone EZ, Weinfurt KP, Friedman JY, Schulman KA, Bosworth HB, Perceived discrimination and reported delay of pharmacy prescriptions and medical tests, J. Gen. Intern. Med 20 (2005) 578–583. 10.1111/j.1525-1497.2005.0123.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Kirk JK, D’Agostino RB, Bell RA, Passmore LV, Bonds DE, Karter AJ, Narayan KMV, Disparities in HbA1c levels between African-American and non-Hispanic white adults with diabetes: A meta-analysis, Diabetes Care. 29 (2006) 2130–2136. 10.2337/dc05-1973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Mikhail GW, Coronary heart disease in women, BMJ. 331 (2005) 467–468. 10.1136/bmj.331.7515.467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Williams RA, Cardiovascular disease in African American women: A health care disparities issue, J. Natl. Med. Assoc 101 (2009) 536–540. 10.1016/S0027-9684(15)30938-X. [DOI] [PubMed] [Google Scholar]
- [32].Piette JD, Schillinger D, Potter MB, Heisler M, Dimensions of patient-provider communication and diabetes self-care in an ethnically diverse population, J. Gen. Intern. Med 18 (2003) 624–633. 10.1046/j.1525-1497.2003.31968.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Schillinger D, Bindman A, Wang F, Stewart A, Piette J, Functional health literacy and the quality of physician-patient communication among diabetes patients, Patient Educ. Couns 52 (2004) 315–323. 10.1016/S0738-3991(03)00107-1. [DOI] [PubMed] [Google Scholar]
- [34].Collins TC, Clark JA, Petersen LA, Kressin NR, Racial differences in how patients perceive physician communication regarding cardiac testing., Med. Care 40 (2002). 10.1097/00005650-200201001-00004. [DOI] [PubMed] [Google Scholar]
- [35].Martin KD, Roter DL, Beach MC, Carson KA, Cooper LA, Physician communication behaviors and trust among black and white patients with hypertension, Med. Care 51 (2013) 151–157. 10.1097/MLR.0b013e31827632a2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Johnson RL, Saha S, Arbelaez JJ, Beach MC, Cooper LA, Racial and ethnic differences in patient perceptions of bias and cultural competence in health care, J. Gen. Intern. Med 19 (2004) 101–110. 10.1111/j.1525-1497.2004.30262.x. [DOI] [PMC free article] [PubMed] [Google Scholar]