Chronic Illness Self-Management Latent Profiles in Individuals With Comorbid Type 2 Diabetes and Hypertension

Abstract

Purpose:

The purpose of the study was to identify chronic illness self-management latent profiles based on health literacy, disease knowledge, self-efficacy, disease distress, perceived self-care, and medication use among individuals with type 2 diabetes and hypertension.

Methods:

This cross-sectional study utilized baseline data from 220 participants in the EXTEND trial. Latent profile analysis was employed to identify distinct chronic illness self-management profiles based on attributes such as health literacy, disease knowledge, self-efficacy, disease distress, perceived self-care, and medication use. Sociodemographic and clinical data were analyzed to explore associations with these profiles.

Results:

Three profiles were identified: “high distress” (18.1%), characterized by high disease distress and a high medication use score, indicating greater inconsistency in medication-taking behaviors; “high health knowledge” (38.6%), with high health literacy and disease knowledge; and “high self-efficacy” (43.3%), exhibiting high self-efficacy and perceived self-care. Significant differences in A1C levels were observed across profiles, with the high distress group showing the highest A1C levels. Racial identity and socioeconomic factors were significantly associated with profile membership.

Discussion:

This study underscores the complexity of chronic illness self-management in individuals with comorbid diabetes and hypertension. Tailored, multifaceted interventions addressing the unique challenges individuals face in managing their illnesses are crucial for improving clinical outcomes and promoting health equity.

Managing comorbid type 2 diabetes (T2DM) and hypertension presents significant physical and psychological challenges. In 2021, 38.1 million US adults (14.7%) had diabetes, and 70.8% of these individuals had elevated blood pressure (BP; ≥140/90 mmHg) or were on medication for high blood pressure.¹ Diabetes and hypertension share risk factors and contribute to cardiovascular diseases and other complications, including infections, cancer, and kidney disease.^2,3 These challenges complicate their care and overall well-being, negatively impacting quality of life. Moreover, the burdens associated with this comorbidity have led to a surge in direct and indirect costs, including hospitalizations and premature mortality.⁴ These burdens necessitate extensive lifestyle modifications, engagement with complex medication regimens, ongoing monitoring, regular medical checkups, and coping with the psychological, financial, and social consequences of chronic illness management.⁵

In the face of challenges in managing T2DM and hypertension, self-management is critical to enhance health outcomes and reduce the dual burden of these chronic conditions. Chronic illness self-management (CISM) involves a lifelong commitment to maintaining prescribed treatments, including medication regimens, dietary restrictions, or physical activity guidelines.^6,7 Furthermore, CISM encompasses the ability to adapt to changes in health status—responding to fluctuations in blood glucose levels or BP readings—and the resilience to preserve one’s overall wellness despite the challenges posed by the illnesses.⁸ In essence, navigating CISM in this population is important in the design of person-centered health care strategies.

Effective CISM requires individuals to possess skills, health knowledge, and positive attitudes to make informed decisions and address illness complexities.⁶ Previous studies reveal that attributes of CISM, such as diabetes knowledge,⁹ health literacy,¹⁰ diabetes distress,¹¹ and consistent medication use,¹² are associated with improved health outcomes in individuals with T2DM. Similarly, in hypertension management, attributes such as health literacy,¹³ disease-specific knowledge,¹⁴ psychological distress,¹⁵ and consistent medication use¹⁶ are associated with BP. These attributes contribute to the effective management of these chronic conditions and underscore the importance of targeted interventions to enhance self-management practices. However, there is limited research on how these attributes combine and influence CISM in individuals with comorbid diabetes and hypertension. Understanding the complex interplay of CISM attributes in this population is crucial for developing comprehensive care models to improve clinical outcomes and quality of life for these individuals.

This study seeks to address this gap by employing latent profile analysis (LPA) to examine latent profiles of CISM among individuals with T2DM and hypertension. LPA enables the exploration of the complex interplay of CISM attributes, identifying distinct, homogeneous subgroups in a population by analyzing observable health behavior variables.^17,18 Unlike traditional variable-centered methods, which focus on the relationships between variables, LPA explores the patterns of attributes, their prevalence, and the background characteristics of individuals exhibiting specific patterns.¹⁹ Therefore, this study seeks to identify CISM profiles based on health literacy, disease knowledge, self-efficacy, disease distress, perceived self-care, and medication use. Specifically, the study has 2 objectives: first, to identify homogeneous profiles characterized by shared attributes in CISM among individuals with comorbid T2DM and hypertension and second, to assess the relationship between the CISM profiles, sociodemographic characteristics, and clinical outcomes, specifically, A1C levels and BP.

The CISM behavior model, developed by Hill-Briggs,⁶ serves as the theoretical foundation for this study. The model highlights the need to recognize personal, environmental, social, and knowledge-based barriers to regimen engagement and develop the individual’s CISM capacity to overcome multilevel challenges in chronic condition management . The model suggests that patient education should go beyond teaching disease-specific knowledge and care techniques to include strategies for problem analysis, responses to challenges, and effective utilization of past experiences for future disease-related challenges.

This study employed 3 key components from the theoretical model central to effective CISM: problem-solving skills, problem-solving orientation, and health knowledge. First, problem-solving skills involve identifying problems, finding solutions, and ensuring they work, which is essential for addressing the complexities associated with chronic conditions. In this study, problem-solving skills are demonstrated through medication use and perceived self-care, which represent practical steps individuals take in managing diabetes and hypertension. Second, problem-solving orientation is operationalized by self-efficacy and disease distress, illustrating how personal attitudes and emotional responses influence diabetes and hypertension self-management. Lastly, health knowledge refers to understanding a disease, its management, and potential outcomes, operationalized through health literacy and disease knowledge.

Methods

Study Design and Participants

This cross-sectional, descriptive study used baseline data from the ongoing EXpanding Technology-Enabled, Nurse-Delivered Chronic Disease Care (EXTEND) trial (ClinicalTrials.gov: NCT05120544). The EXTEND study is a randomized trial assessing two 12-month interventions for individuals with T2DM and hypertension: one involving mobile monitoring for self-management and the other a nurse-led telehealth program that includes mobile monitoring, self-management support, and pharmacist-assisted medication management.^20,21 The full protocol, baseline data, and technology build for the EXTEND trial have been previously reported.^20,21 The current study focuses on baseline cognitive, behavioral, and psychological characteristics related to CISM, providing new insights into self-management characteristics.

Eligibility for the EXTEND trial was defined by the following criteria: participants ages 30 to 75, diagnosed with T2DM and hypertension per International Classification of Diseases codes; a history of diabetes, indicated by a consistently high A1C level (≥8.0%) for at least 6 months with no readings below this threshold; a hypertension diagnosis supported by a clinic BP reading over 140 systolic or 90 diastolic within the last year; and ownership of a smartphone. A total of 220 individuals met these criteria, provided informed consent, and completed a baseline assessment. The Duke University Health System Institutional Review Board approved the trial (Pro00107722). The current study maintains the initial study’s eligibility criteria as the standard for inclusion in the current investigation.

Measures

The study gathered sociodemographic data through self-report measures and included variables such as age, gender, race, marital status, educational attainment, employment status, and annual household income. Race was categorized as Black or African American, White, or other (Asian, American Indian, Alaska Native, or more than one race) because of the small sample sizes in the other categories. In addition to sociodemographic data, the study examined 3 baseline clinical outcomes: (1) average systolic BP, (2) average diastolic BP, and (3) A1C level. BP was measured twice to determine the average systolic and diastolic BP readings in millimeters of mercury (mmHg). The A1C level (%) was obtained from a single assessment.

Health literacy.

The Newest Vital Sign²² includes 6 items designed to assess an individual’s ability to read and use health-related information. The total score is calculated based on the number of correct responses, with possible scores ranging from 0 to 6; higher scores indicate better health literacy. The scale has demonstrated acceptable internal consistency in prior research (Cronbach’s α = 0.75).²³ Based on the total sample of participants in this study, the scale showed acceptable internal consistency, with a Cronbach’s α of 0.75.

Disease knowledge.

The Diabetes Knowledge Questionnaire (DKQ-24)²⁴ and Hypertension Knowledge Measure,¹⁴ consisting of a combined total of 24 items, were employed to evaluate individuals’ knowledge about diabetes and hypertension. Responses were structured as “yes,” “no,” or “I don’t know,” with each response being scored as correct (1) or incorrect (0). The total score was calculated based on the number of correct answers, with possible scores ranging from 0 to 24. Previous studies reported acceptable internal consistency for the DKQ-24 (Cronbach’s α = 0.78)²⁴ and the Hypertension Knowledge Measure (Cronbach’s α = 0.70).¹⁴ In this study, the combined measure demonstrated acceptable internal consistency (Cronbach’s α = 0.77).

Self-efficacy.

The Perceived Competence Scale is a 4-item scale that measures individuals’ self-efficacy in managing daily self-management tasks.²⁵ Each item is rated on a scale from 1 (not at all true) to 7 (very true) to determine the extent of agreement. The average of these item scores forms the overall scale score, with possible scores ranging from 1 to 7, where higher values indicate greater self-efficacy in self-management. In this study, the scale demonstrated good internal consistency, with a Cronbach’s α of 0.83.

Perceived self-care.

Participants completed the 16-item Diabetes Self-Management Questionaire,^26,27 which evaluated their self-assessed engagement with recommended self-care practices over the past month. Responses were measured using a 5-point Likert scale from 1 (never do it) to 5 (always do this as recommended without fail). The overall score was derived by averaging the responses and scaling them to a 0- to 10-point range, with higher scores representing greater perceived self-care. The scale demonstrated good internal consistency in prior research (Cronbach’s α = 0.84)²⁶ and in this study (Cronbach’s α = .89).

Medication use.

Diabetes and BP medication use was assessed using the Medication Non-Adherence Scale.²⁸ This scale includes 6 items, with 3 items each for diabetes and BP medications. The scale employs a 6-point Likert scale to assess the frequency of patterns in medication taking, from 1 (none of the time) to 6 (every time). The overall score is the mean of the item-level scores, ranging from 1 to 6, where higher scores reflect less frequent medication use or greater inconsistency in medication taking. The scale has demonstrated good internal consistency in prior research (Cronbach’s α = 0.84)²⁸ and in this study (Cronbach’s α = 0.86).

Disease distress.

Disease distress was assessed using a 17-item modified version of the Diabetes Distress Scale, which included additional questions addressing hypertension.^29,30 The measure uses a 6-point Likert scale, ranging from 1 (not a problem) to 6 (a very serious problem). The overall score was the average of the item-level scores; higher scores indicate greater disease distress. The original scale has shown good internal consistency in prior research (Cronbach’s α = 0.87).³⁰ In this study, the modified scale showed excellent internal consistency (Cronbach’s α = 0.93) within the total sample.

Statistical Analyses

A descriptive analysis was conducted to summarize the individual characteristics and clinical outcomes. LPA was performed using the “mclust” package in R software version 4.4.0. LPA is used to identify unobserved (latent) subgroups in a population based on patterns in individuals’ responses to a set of variables.¹⁸ Instead of assuming a homogeneous population, LPA classifies individuals into distinct profiles that share similar characteristics, providing a clearer understanding of variability in the sample.¹⁹ The final model selection was based on goodness-of-fit indices, simplicity, and clinical interpretability to ensure accurate representation of CISM profiles. Model fit indices evaluated how well each latent profile solution represented the data and helped determine the optimal number of profiles. These included the log-likelihood, Akaike information criterion (AIC), Bayesian information criterion (BIC), consistent Akaike information criterion (CAIC), and sample-size adjusted Bayesian information criterion (SABIC). Lower values for these indices indicate a better fitting model, suggesting that the chosen model captures meaningful subgroups while balancing model complexity.¹⁷ Entropy, a standardized classification accuracy measure ranging from 0 to 1, was used to evaluate the precision in assigning individuals to profiles, with higher values suggesting greater accuracy.^17,18

To explore differences in sociodemographic and clinical characteristics associated with different CISM latent profiles, chi-square tests were conducted for categorical variables, and analysis of variance was employed for continuous variables. Tukey’s honestly significant difference test was used for post hoc comparisons. To further characterize the associations, multinomial logistic regression was used to calculate the odds ratio (OR) for being in one profile compared to a reference level for each of the characteristics. Given the exploratory nature of this study, a threshold P value of less than .05 was established to denote statistical significance. All tests were conducted non-directionally at this level of significance.

Results

Table 1 presents demographics, CISM characteristics, and clinical outcomes for 220 participants. The average age of participants was 54.5 years (SD = 10.3). The majority of the sample were female (63.6%), and most identified as Black or African American (68.2%). Educational levels varied, with 25.0% having completed high school or less and 33.6% possessing a university or college degree. In terms of CISM attributes, the mean health literacy score was 3.1 (SD = 2.0), and the mean score for medication use was 1.9 (SD = 0.9). For the clinical outcomes, the average A1C level was 9.8 (SD = 1.7).

Table 1.

Individual Sociodemographic, Chronic Illness Self-Management, and Clinical Characteristics (N = 220).

Variables	Overall (N = 220)
Age, mean (SD)	54.5 (10.3)
Gender, n (%)
Male	79 (35.9)
Female	140 (63.6)
Decline to answer or missing	1 (0.5)
Race, n (%)
Black or African American	150 (68.2)
White	47 (21.4)
Other	9 (4.1)
Decline to answer or missing	14 (6.3)
Married/partnered, n (%)
Yes	101 (46.0)
No	114 (51.8)
Decline to answer or missing	5 (2.2)
Full-/part-time employment, n (%)
Yes	122 (55.4)
No	89 (40.5)
Decline to answer or missing	9 (4.1)
Highest education attainment, n (%)
High school or less	55 (25.0)
Postsecondary education/traininga	87 (39.6)
University/college degree	74 (33.6)
Decline to answer or missing	4 (1.8)
Annual household income, n (%)
Less than $10 000	37 (16.8)
$10 000 to $29.9 000	49 (22.3)
$30 000 to $49.9 000	51 (23.2)
$50 000 to $79.9 000	33 (15.0)
$80 000 or greater	22 (10.0)
Decline to answer or missing	28 (12.7)
Chronic illness self-management attributes [Scale range], mean (SD)
Health literacy [0-6]	3.1 (2.0)
Disease knowledge [0-24]	16.3 (3.8)
Self-efficacy [1-7]	4.8 (1.5)
Perceived self-care [0-10]	6.2 (1.5)
Medication use [1-6]	1.9 (0.9)
Disease distress [1-6]	2.5 (1.0)
Clinical outcomes, Mean (SD)
Systolic BP	134.9 (20.2)
Diastolic BP	81.1 (9.3)
A1C	9.8 (1.7)

Abbreviations: A1C, A1C level in percentage; BP, blood pressure (mmHg).

^aPostsecondary education/training: some formal education or training following high school graduation.

Model Selection and Fit Analysis for Chronic Illness Self-Management Profiles

Table 2 presents a detailed summary of the model fit indices for each profile. The BIC and CAIC were the lowest for the 3-latent-class model. Additionally, the log-likelihood, AIC, and SABIC continued to decrease with the addition of more latent profiles, but a leveling off was observed after the 3-latent-profile solution. Further examination of models with 4 to 8 latent profiles revealed issues with homogeneity and latent profile separation, which were not as pronounced in the 3-profile solution. Although entropy indicated that models with 6 to 8 profiles might be preferable, the overall balance of fit and parsimony favored the 3-latent-profile model. This study selected the 3-latent-profile model based on the comprehensive analysis of fit indices, parsimony, and the interpretative quality of the resultant latent classes.

Table 2.

Latent Profile Model Fit Indices for 1 to 8 Profiles

No. of profiles	Model fit indices
	LogLik	AIC	BIC	CAIC	SABIC	Entropy
1	−1870	3764	3805	3817	3767	1
2	−1797	3632	3696	3715	3636	0.825
3	−1773	3596	3684	3710	3601	0.728
4	−1760	3586	3698	3731	3593	0.685
5	−1743	3567	3703	3743	3576	0.757
6	−1722	3537	3697	3744	3548	0.784
7	−1705	3518	3701	3755	3530	0.793
8	−1700	3522	3729	3790	3536	0.803

Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion; CAIC, consistent Akaike information criterion; LogLik, log-likelihood; SABIC, sample-size adjusted Bayesian information criterion.

Chronic Illness Self-Management Attributes Across the Profiles

Table 3 presents unstandardized mean scores, and Figure 1 shows the distribution of standardized mean scores of CISM attributes across the 3 latent profiles. Profile 1 (n = 40, 18.1%) was characterized by the highest levels of medication use and disease distress coupled with the lowest levels of self-efficacy and perceived self-care. Profile 1 was labeled as having “high distress.” Profile 2 (n = 85, 38.6%) featured high health literacy and disease knowledge. This profile also showed low disease distress and higher perceived self-care. Consequently, Profile 2 was labeled as “high health knowledge.” Profile 3 (n = 95, 43.3%) showed the lowest health literacy and disease knowledge but the highest self-efficacy and perceived self-care. Thus, Profile 3 is labeled as “high self-efficacy.”

Table 3.

Sociodemographic and Clinical Characteristics Across the Profiles

Variables	Latent profiles			Test statistics (df)	P
	1. High distress	2. High health knowledge	3. High self-efficacy
Clustering size, n (%)	40 (18.1)	85 (38.6)	95 (43.3)
Chronic illness self-management attributes [scale range], mean (SD)
Health literacy [0-6]	2.8 (1.9)	5.0 (1.0)	1.7 (1.1)
Disease knowledge [0-24]	14.9 (4.0)	18.6 (2.5)	14.8 (3.7)
Self-efficacy [1-7]	3.6 (1.7)	4.7 (1.0)	5.4 (1.4)
Perceived self-care [0-10]	4.4 (1.2)	6.1 (1.3)	7.0 (1.1)
Medication use [1-6]	2.9 (0.9)	1.7 (0.7)	1.7 (0.7)
Disease distress [1-6]	3.9 (1.0)	2.2 (0.7)	2.1 (0.8)
Sociodemographic characteristics
Gender, n (%)				χ2 = 2.75 (2)a	.253
Male	10 (25.0)	34 (40.0)	35 (36.8)
Female	30 (75.0)	51 (60.0)	59 (62.1)
Decline to answer or missing	0 (0.0)	0 (0.0)	1 (1.05)
Race, n (%)				χ2=16.90 (4)a	.002
Black or African American	34 (85.0)	54 (63.5)	62 (65.3)
White	1 (2.5)	28 (32.9)	18 (18.9)
Other	2 (5.0)	1 (1.2)	6 (6.3)
Decline to answer or missing	3 (7.5)	2 (2.4)	9 (9.5)
Married/partnered, n (%)				χ2 = 10.50 (2)a	.005
Yes	28 (70.0)	25 (41.2)	51 (53.7)
No	10 (25.0)	48 (56.5)	43 (45.3)
Decline to answer or missing	2 (5.0)	2 (2.3)	1 (1.0)
Full-/part-time employment, n (%)				χ2 = 7.70 (2)a	.021
Yes	10 (25.0)	31 (36.5)	48 (50.5)
No	28 (70.0)	49 (57.6)	45 (47.4)
Decline to answer or missing	2 (5.0)	5 (5.9)	2 (2.1)
Highest education attainment, n (%)				χ2 = 13.92 (4)a	.008
High school or less	10 (25.0)	13 (15.3)	32 (33.7)
Postsecondary education/training	18 (45.0)	31 (36.5)	38 (40.0)
University/college degree	9 (22.5)	40 (47.0)	25 (26.3)
Decline to answer or missing	3 (7.5)	1 (1.2)	0 (0.0)
Annual household income, n (%)				χ2 = 22.09 (8)a	.005
Less than $10 000	9 (22.5)	10 (11.8)	18 (19.0)
$10 000 to $29.9 000	7 (17.5)	13 (15.3)	29 (30.5)
$30 000 to $49.9 000	10 (25.0)	18 (21.2)	23 (24.2)
$50 000 to $79.9 000	5 (12.5)	15 (17.6)	13 (13.7)
$80 000 or greater	1 (2.5)	17 (20.0)	4 (4.2)
Decline to answer or missing	8 (20.0)	12 (14.1)	8 (8.4)
Clinical outcomes, mean (SD)
Systolic BP	137.5 (20.5)	133.0 (18.1)	135.5 (21.8)	F = 0.74 (2)b	.479
Diastolic BP	83.9 (10.5)	79.9 (8.5)	80.9 (9.2)	F = 2.57 (2)b	.098
A1C	10.7 (1.7)	9.5 (1.6)	9.7 (1.7)	F = 7.60 (2)b	<.001

Note: P values in bold indicate statistical significance at P < .05.

Abbreviations: A1C, A1C level in percentage; BP, blood pressure (mmHg).

^aChi-square test.

^bAnalysis of variance.

Figure 1.

Distribution of standardized mean scores of chronic illness self-management attributes across 3 latent profiles. The figure illustrates the standardized mean scores of 6 chronic illness self-management attributes—health literacy, disease knowledge, self-efficacy, perceived self-care, disease distress, and medication use—across 3 identified profiles: “high distress” (Profile 1), “high health knowledge” (Profile 2), and “high self-efficacy” (Profile 3). Higher scores for medication use indicate greater inconsistency in medication-taking behaviors.

Sociodemographic Characteristics and Clinical Outcomes Across the Profiles

Table 3 presents the distribution of sociodemographic characteristics and clinical outcomes across the 3 distinct profiles. Significant differences in racial distribution were identified (χ² = 16.90, P = .002), with Black or African American individuals being predominant in Profile 1 (high distress). Educational attainment varied significantly among the groups (χ² = 13.92, P = .008). Profile 2 (high health knowledge) had the highest proportion of individuals holding a university or college degree, whereas Profile 3 (high self-efficacy) had the greatest representation of participants with a high school diploma or less.

For the clinical outcomes, no significant differences were observed in systolic and diastolic BP among the profiles. However, significant disparities in A1C levels were observed across the profiles (F = 7.60, P < .001). Profile 1 had the highest mean A1C level, averaging 10.7 (SD = 1.7). Tukey’s post hoc test confirmed these findings, showing that the mean A1C level for Profile 1 significantly differed from those of Profiles 2 and 3, with mean differences of 1.33 (95% CI, 0.59-2.07) and 1.18 (95% CI, 0.45-1.91), respectively (both P < .001).

Associations Between Sociodemographic Characteristics, Clinical Outcomes, and CISM Profiles

The associations between the explanatory variables and the latent profiles are presented in Table 4. Participants who identified as White were 94% less likely to be in Profile 1 compared to Profile 2 (OR = 0.06, 95% CI, 0.01-0.42, P = .005). Additionally, participants with an annual household income of $80 000 or greater were 87% less likely to be in Profile 1 (OR = 0.13, 95% CI, 0.02-0.73, P = .020) and 90% less likely to be in Profile 3 (OR = 0.10, 95% CI, 0.02-0.42, P = .002) compared to Profile 2. Higher A1C levels were associated with an increased likelihood of being in Profile 1 compared to Profile 2, with an OR of 1.55 (95% CI, 1.25-1.93, P < .001).

Table 4.

Associations Between Sociodemographic Variables, Clinical Outcomes, and Chronic Illness Self-Management Profiles

Variable	Profile 1 vs Profile 2 (reference)			Profile 3 vs Profile 2 (reference)
	OR	95% CI	P	OR	95% CI	P
Gender (reference = male)
Female	1.81	0.80-4.07	.153	1.06	0.58-1.94	.859
Race (reference = Black or African American)
White	0.06	0.01-0.42	.005	0.57	0.28-1.14	.113
Other	1.03	0.16-6.47	.976	1.18	0.25-5.51	.833
Marital status (reference = not married, partnered)
Married/partnered	0.27	0.12-0.60	.001	0.56	0.31-1.02	.058
Work status (reference = no full-/part-time employment)
Full-/part-time employment	1.92	0.83-4.46	.129	0.57	0.31-1.05	.072
Highest education attainment (reference = high school or less)
Postsecondary education/training	0.77	0.28-2.10	.612	0.46	0.21-1.02	.056
University/college degree	0.32	0.11-0.93	.036	0.23	0.10-0.52	<.001
Annual household income (reference = less than $10 000)
$10 000 to $29.9 000	0.56	0.15-1.99	.367	1.11	0.41-3.03	.837
$30 000 to $49.9 000	0.58	0.18-1.91	.374	0.64	0.24-1.73	.381
$50 000 to $79.9 000	0.44	0.12-1.64	.224	0.44	0.15-1.31	.142
$80 000 or greater	0.13	0.02-0.73	.020	0.10	0.02-0.42	.002
Clinical outcomes
Systolic BP	1.01	0.99-1.03	.346	1.00	0.99-1.02	.517
Diastolic BP	1.04	0.36-1.09	.031	1.02	0.98-1.05	.316
A1C	1.55	1.25-1.93	<.001	1.11	0.92-1.35	.277

Note: P values in bold indicate statistical significance at P < .05.

Abbreviations: A1C, A1C level in percentage; BP, blood pressure (mmHg); OR, odds ratio.

Discussion

This study identified 3 latent profiles of CISM among individuals with comorbid T2DM and hypertension: Profile 1 (high distress), Profile 2 (high health knowledge), and Profile 3 (high self-efficacy). The findings revealed unique patterns of CISM attributes, underscoring the complexity of CISM in this population. Individuals in Profile 1 are characterized by high levels of distress and scores for medication use, indicating heightened psychological challenges and greater inconsistency in medication-taking behaviors. Profile 2 is associated with greater health knowledge, which may enhance individuals’ ability to navigate chronic conditions more effectively.^10,31 In contrast, Profile 3 highlights the importance of self-efficacy in sustaining effective self-management. This result suggests different challenges individuals might face in managing their chronic conditions, highlighting the necessity for a multifaceted care model that addresses the distinct challenges.

The results revealed significant differences in A1C levels across the latent profiles. Profile 1 exhibited the highest A1C levels. Profile 2, which demonstrated good CISM attributes, showed the lowest mean A1C level. Notably, no significant difference in A1C levels was found between Profile 2 and Profile 3. Despite having the highest perceived self-care and self-efficacy, Profile 3 exhibited the lowest health literacy and disease knowledge scores. The similar A1C levels in Profiles 2 and 3, despite their large differences in health knowledge, suggest that the perceived self-care and self-efficacy in Profile 3 might be critically related to effective blood glucose management. However, this finding necessitates caution in interpreting the results because it does not imply that Profiles 2 and 3 are proficient in managing blood glucose but, rather, that these individuals are relatively better managed compared to Profile 1. These findings highlight the need to investigate the mechanisms that enable individuals in Profile 3 to maintain their A1C with lower health knowledge because understanding this could lead to effective and targeted clinical interventions for diabetes management.

The observed differences in A1C levels across CISM profiles underscore the importance of psychological attributes and individuals’ perceptions of their self-management. Distress in managing chronic conditions is associated with less favorable health outcomes, further increasing the overall burden.³ Conversely, self-efficacy and perceived self-care contribute to better health outcomes by enabling proactive CISM behaviors.^32,33 High levels of perceived self-care and self-efficacy indicate that individuals feel capable and motivated in managing their health. Enhancing competence is crucial for improving health outcomes and reducing the burdens of chronic illnesses such as diabetes and hypertension by bolstering intrinsic motivation and addressing unique self-management challenges.

Sociodemographic characteristics varied across the profiles, elucidating the societal and structural challenges faced in managing chronic illnesses. The results highlight the complex interplay between socioeconomic factors and health outcomes. For instance, education level showed a statistically significant association with the latent profiles, with those possessing university or college degrees being less likely to fall into Profile 1. Moreover, individuals with a yearly household income of $80 000 or greater were less likely to be in Profile 1 compared to Profile 2. These findings suggest a significant impact of socioeconomic factors on clinical outcomes given that higher A1C levels were more prevalent in Profile 1. Additionally, individuals who identified as Black or African American were more likely to belong to Profile 1 than individuals who identified as White. This finding underscores the need to address the challenges faced by different racial groups. The disproportionate influence of chronic illness on racial and ethnic minority groups is critically associated with structural racism, which leads to significant health inequities, including systemic barriers such as limited access to health care and lack of culturally tailored health education and support programs.^34-36 Therefore, health care models for promoting diabetes and hypertension CISM must be tailored to recognize and address these challenges by understanding the systemic barriers and support networks that individuals and communities use.^37,38 Incorporating a deeper understanding of these contextual dynamics can inform the development of more inclusive strategies that promote equity in chronic illness management.

Applying a person-centered approach through latent modeling, previous studies have primarily identified CISM patterns based on specific behaviors. For instance, Shen et al³⁹ classified individuals with diabetes according to their engagement with medication regimen, dietary practices, physical activity, and glucose monitoring, observing better glucose management in those consistently maintaining these behaviors. Similarly, Nam and Yoon⁴⁰ explored hypertension management, identifying groups with varying levels of engagement in smoking cessation, diet, and exercise. Heise et al⁴¹ examined diabetes self-management patterns, including dietary planning and blood glucose monitoring, and highlighted the significant impact of structured diabetes education programs on improving self-management engagement. These studies underscore the need for interventions tailored to individuals’ unique self-management behaviors. Tailored support can benefit individuals with chronic conditions by equipping them with practical skills and enhancing engagement in CISM. Although these studies provide valuable insights, they often overlook dynamic attributes that influence self-management. In contrast, this study broadens the perspective on CISM by incorporating problem-solving skills, orientation, and health knowledge as central elements. Focusing on psychological and cognitive attributes and adaptive skills allows for a shift beyond behavior-focused strategies, offering a more comprehensive foundation for intervention design. This broader perspective may contribute to greater engagement and foster sustainable health outcomes by enhancing individuals’ capacity for CISM in diabetes and hypertension.

This study has several limitations. First, the cross-sectional design limits the ability to conclude causal relationships between profile membership and clinical outcomes. Future research should aim to address this limitation by employing a longitudinal design. A longitudinal design would enable observation of how changes in CISM attributes influence health outcomes over time, offering a dynamic understanding of profile membership, causal pathways, and long-term impacts on health. Second, the probabilistic classification of participants into latent profiles introduces uncertainty because some may not be perfectly classified, potentially affecting profile accuracy. Third, the sample size and its demographic composition can limit the generalizability of the results. There are differing opinions on the optimal sample size for LPA, with suggestions ranging from 200 to 500 or more.^42-44 Although this study, which consists of 220 individuals, provides valuable insights, the specific socioeconomic and demographic backgrounds of participants from North Carolina, USA, may not fully capture the diversity of the broader population of individuals with diabetes and hypertension. Therefore, this study highlights the need for caution when extrapolating these results to other populations.

Furthermore, the theoretical model employed in this study emphasizes individual-level factors as key factors of self-management success. However, this approach does not fully account for broader social determinants of health, including access to health care, socioeconomic status, and systemic barriers.³⁷ These factors significantly influence both unintentional and intentional CISM engagement.⁴⁵ Individuals’ choices are often constrained by their circumstances, and engagement patterns can reflect adaptive strategies to navigate limited resources rather than a simple lack of problem-solving ability. Therefore, further research should incorporate models that encompass a more comprehensive understanding of self-management, including the impact of social determinants of health. This would provide a holistic perspective on the intersections of individual and systemic factors, such as access to health care, trust in the system, and socioeconomic disparities, to better capture the complexity of CISM.

Conclusion

CISM is crucial for improving health outcomes, enhancing quality of life, and ensuring lifelong commitment to health maintenance. By employing LPA, this study elucidates the distinct patterns of CISM attributes in individuals with T2DM and hypertension. The findings highlight the challenges individuals face in managing their chronic illnesses. By acknowledging the unique characteristics and challenges, this study informs the design of multifaceted comprehensive care models that consider broader social determinants of health. Incorporating a deeper understanding of these contextual dynamics can inform the development of effective strategies that improve self-management behaviors and enhance health outcomes. This approach will contribute to reducing health disparities and promoting equity in chronic illness management.