H-type hypertension, defined in the study as hypertension accompanied by plasma homocysteine of at least 10 μmol/L, combines elevated blood pressure with hyperhomocysteinemia. The authors noted that the two abnormalities together are associated with a greater risk of stroke than either elevated blood pressure or homocysteine alone, while high homocysteine has also been linked with endothelial dysfunction, vascular oxidative stress, arterial stiffening, coronary artery disease, heart failure and the progression of chronic kidney disease. The findings concern a population for which sustained patient-led management may be particularly important, although the study did not test whether any self-management pattern reduced cardiovascular events or improved long-term blood-pressure control.

Participants were recruited by convenience sampling from the outpatient department and health-examination center of one tertiary hospital in Shandong Province between June and December 2024. Eligible adults were at least 18 years old, had H-type hypertension for at least six months and provided informed consent. Patients with secondary hypertension or serious conditions that could limit the exercise, dietary and monitoring behaviors assessed by the questionnaire were excluded, including New York Heart Association class III–IV heart failure, recent acute coronary syndrome, disabling residual stroke, advanced chronic kidney disease or dialysis, oxygen-dependent chronic obstructive pulmonary disease and active malignancy. Of 323 returned paper questionnaires, 13 were excluded for incomplete or patterned responses, leaving 310 complete cases for analysis.

The investigators assessed six dimensions of behavior with a 33-item hypertension self-management scale covering medication adherence, self-monitoring, diet, exercise, activity and rest, and emotion regulation. Mean scores for these dimensions ranged from 2.77 out of 5 for exercise management to 3.94 for medication adherence. The standardized overall self-management score was 67.34 ± 13.82 on a theoretical scale of 20 to 100, which the researchers categorized as moderate. That aggregate figure nevertheless combined patients with very different behavioral configurations. Latent profile analysis of the six dimension scores identified the four groups, with an entropy of 0.883 and average posterior probabilities ranging from 0.878 to 0.999. The profile labels describe multidimensional patterns and do not correspond directly to the scale’s low, moderate and high overall-score thresholds.

The largest group, the high self-management profile, accounted for 41.6% of the sample and had the highest scores across all six dimensions, although exercise remained a relative weakness. The low self-management group, representing 18.4%, had the lowest scores throughout and, in unadjusted comparisons, the lowest eHealth literacy, perceived social support and psychological resilience as well as the highest perceived stress. Patients in the moderate treatment-monitoring profile, 21.0% of the sample, scored relatively better on lifestyle behaviors but less well on medication adherence and self-monitoring. The lifestyle-focused moderate group, comprising 19.0%, retained comparatively stronger medication adherence and monitoring but had lower scores for exercise, activity and rest, and emotion regulation.

Multinomial logistic regression used the high self-management profile as its reference category and accounted for uncertainty in class assignment through the R3STEP method. Compared with patients without comorbidity, those with one or more of the included, relatively mild comorbidities had lower odds of belonging to the low self-management group, with an odds ratio of 0.067 and a 95% confidence interval of 0.005 to 0.889, and to the moderate treatment-monitoring group, with an odds ratio of 0.240 and a 95% confidence interval of 0.096 to 0.602. The association with the moderate lifestyle profile was not statistically significant. Because the study was cross-sectional and lacked measures of healthcare contact, treatment burden and health awareness, the authors could not determine why comorbidity showed this relationship.

Higher eHealth literacy was associated with a lower likelihood of membership in the moderate lifestyle profile rather than the high self-management group, with an odds ratio of 0.927 per additional point and a 95% confidence interval of 0.890 to 0.966. Perceived social support showed a similar association for both moderate groups: the odds ratios were 0.951 for the treatment-monitoring comparison and 0.942 for the lifestyle comparison, with respective 95% confidence intervals of 0.911 to 0.993 and 0.899 to 0.987. Its association with the low profile was not significant. Hypertension duration, hospital anxiety, occupation and medical-insurance type also failed to reach statistical significance in the final model. The null findings do not demonstrate that these distal social or structural factors are unimportant, particularly because the smallest group contained 57 participants and the model had about 7.1 events per parameter.

The results suggest that a single “moderate” average conceals clinically distinct needs. Patients in the treatment-monitoring profile may require support for taking medicines and recording relevant measures, whereas those in the lifestyle profile may have greater needs around exercise, rest and emotion regulation. The study did not measure the reasons for these weaknesses, and the authors cautioned that the treatment-monitoring label should not be interpreted as a validated H-type-specific phenotype: the instrument did not directly assess folic acid or B-vitamin use, dietary folate or homocysteine retesting. The findings therefore support further development and prospective evaluation of profile-sensitive strategies, but they do not show that tailoring care to a profile improves outcomes.

Several limitations constrain the interpretation and potential reach of the results. Recruitment from one hospital and the high proportion of participants using tertiary hospitals as their usual source of care mean the profiles may differ from those in primary care and rural communities. Excluding serious comorbidities likely produced a healthier sample than the wider H-type hypertension population, while self-reported behavior was vulnerable to recall and social-desirability bias. A single baseline blood-pressure reading could not establish sustained control, and data on medication regimens, treatment burden, objective adherence, renal function and longitudinal cardiovascular outcomes were unavailable. The profiles were not externally validated, their prognostic value remains unknown, and the sample was below the more conservative size of approximately 500 sometimes recommended for reliable class enumeration. Multicenter, community-based and longitudinal studies are therefore needed to test whether the four patterns persist, predict clinical endpoints and meaningfully guide individualized care.

Subject of Research: Latent profiles and associated factors in self-management behaviors among adults with H-type hypertension

Article Title: Self-management behaviors in patients with H-type hypertension: a latent profile analysis within the health ecological model

Article References: Ma, L., Liang, J., Fang, X., Yang, M., Zhang, H., Zhao, T., & Liu, T. (2026). Self-management behaviors in patients with H-type hypertension: a latent profile analysis within the health ecological model. Frontiers in Cardiovascular Medicine, 13, Article 1903891. https://doi.org/10.3389/fcvm.2026.1903891

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1903891

Keywords: H-type hypertension, self-management, latent profile analysis, homocysteine, eHealth literacy, social support, health ecological model, cross-sectional study, patient behavior, hypertension, behaviors, patients