Role of social and other determinants of health in the effect of a multicomponent integrated care strategy on type 2 diabetes mellitus

Diabetes mellitus is a chronic metabolic disease with disabling, deadly, and costly consequences for individuals, families, and national health care systems. Proper diabetes management in people living with type 2 diabetes mellitus (T2DM) focuses on reducing the risks for macrovascular and microvascular complications by controlling blood pressure, lipid levels, and blood glucose levels, and by avoiding tobacco [1]. Although important advances in pharmacological and non-pharmacological strategies have been developed, large gaps exist in achieving care goals, particularly in real-world practice and low- and middle- income countries (LMICs) [2‐4]. Barriers to achieving care goals in T2DM patients may include lack of medical care, poverty, long commutes or lack of time to seek medical attention, lack of confidence, and inadequate social support, among others [5]. Achieving diabetes care goals (e.g., the triple-target goal [6]: normal glycated hemoglobin, blood pressure, and LDL cholesterol) is associated with better health outcomes, including lower risks of complicating events and death [6, 7].

Social determinants of health (SDH), including education level and socioeconomic status (SES), as well as other key factors (diabetes knowledge and self-care scores), named here as other determinants of health (ODH), have demonstrated to modify the response to therapy in T2DM patients [8‐10]. In a previous study, we reported the baseline determinants of glycemic control in T2DM patients and public primary care in Mexico City [11]. It has been suggested that in LMICs urban settings, where the burden of the disease is particularly high, urgent improvements in health care systems are needed to reduce complications and mortality [12, 13]. Diabetes multicomponent programs include strategies for quality on diabetes care improvement directed to patients, health care professionals or systems, particularly focused on enabling periodic evaluations of quality-of-care indicators to identify treatment gaps and disseminate information to promote better decision-making processes [13]. Previous research has shown important improvement in achieving diabetes care goals through multicomponent quality improvement strategies [13, 14]. However, real-world evidence of primary care experiences in public health systems in LMICs is scant for research purposes [13, 15]. The evidence is even scanter for the study of the effect of SDH and ODH on achieving diabetes care goals [15].

Considering these gaps, the purposes of this study were (1) to describe the effect of a diabetes multicomponent integrated care (MIC) program on diabetes care goals (glycated hemoglobin, blood pressure, and LDL cholesterol) and (2) to examine the effect of SDH (education level and SES) and ODH (diabetes knowledge, self-care scores) on T2DM. We aimed to determine the role of SDH and ODH in the context of a multi-intervention program (DIABEMPIC, DIABetes EMPowerment and Improvement of Care), using a before-and-after design and testing the effect of baseline SDH and ODH in the delta (Δ) of change in HbA1c levels.

Mexico City (MC) is one of the largest cities in the world; it is highly polluted and has a high prevalence of obesity, T2DM, and metabolic syndrome. Although ethnically homogeneous, MC also shows wide socioeconomic disparities, with more than 680,000 people living under extreme poverty [16], having low levels of education, lacking health insurance, and constantly exposed to social and domestic violence. Therefore, MC is a strategic place to determine the effects of SDH and ODH on T2DM by using pharmacologic and non-pharmacologic interventions such as DIABEMPIC.

All T2DM patients came from 32 primary outpatient health care centers located in urban areas of Mexico City and were referred to the Clínica Especializada en Diabetes CDMX/Iztapalapa (Specialist Diabetes Clinic) between January 2017 and July 2018. They were invited to participate in the 5-month DIABEMPIC (DIABetes EMPowerment and Improvement of Care) program, which is a primary care strategy of the public primary health care system, designed to improve clinical outcomes in T2DM patients through interdisciplinary care and self-management education schemes. The participation criteria were as follows: (a) T2DM patients older than 18 years, (b) without any acute or chronic complication that required short-term hospital care, and (c) acceptance to participate after understanding the program.

This study aimed to analyze the effect of a MIC intervention on diabetes care goals and to explore potential effects of SDH and ODH on T2DM patients with adverse social characteristics in an urban primary care setting. The DIABEMPIC program significantly improved glycemic control, blood pressure, LDL-C, weight, and HRQOL; it also helped to achieve the diabetes care targets, both separately and combined. We also found that the positive impact for glycemic control was greater among T2DM patients with the lowest education levels and those with the lowest baseline diabetes knowledge and self-care scores, but also among those with greater improvements of diabetes knowledge and self-care activities scores after the intervention. To the best of our knowledge, this is the first study that explores the effect of a MIC intervention and addresses multiple barriers to care in T2DM patients with a predominance of low schooling and low socioeconomic status in a public primary health care system in Mexico. Additionally, to our knowledge, this is one of the largest studies that tries to elucidate the role and contribution of SDH and ODH and clinical determinants of health in the context of a quality improvement intervention in diabetes care goals in an LMIC. A relevant fact is that all the included patients in this study were beneficiaries of Seguro Popular (now INSABI), the national health care system attending the largest proportion of primary care outpatients in Mexico [25]. These patients belong to a population that traditionally lacks formal social and health insurance.

The results of glycemic control shown in this study are more substantial than previously reported [13‐15]. This size effect could be explained by some potential factors: it was previously reported that MIC programs improved clinical outcomes, particularly in young T2DM patients, with suboptimal control and in low-resource settings [13]. The DIABEMPIC program included multiple and different strategies that have demonstrated positive results in diabetes outcomes and surrogate endpoints, all of them integrated into one place. The integration of multiple strategies into an organized way is difficult to occur in real-world practice because it involves health care systems, health care professionals, and T2DM patients. We consider that we have explored the effect of a quality improvement strategy in T2DM patients who had participated in few- or none- of the strategies included in the program, partly due to social backwardness. Thus, we are observing a multiplier effect in a population that is almost unaware of such strategies. To support this factor, we previously reported that poor education levels were linked to poor diabetes knowledge and poor glycemic control [11].

In this study, we observed that the greatest improvements in glycemic control occurred in T2DM patients with the greatest improvements in self-care activities performance and diabetes knowledge. Our results strengthen the recommendation to integrate structured therapeutic education programs linked to improving quality strategies because their impact could be even greater on socially disadvantaged populations. Poverty influences the development of type 2 diabetes and its complications [26]; education and socioeconomic levels are associated with activating self-care management in chronic diseases [27]. Thus, multicomponent quality improvements of care favor narrowing health and social gaps in T2DM patients, as well as some lagging indicators such as poor education level and health education. An even more important fact is the possibility to improve long-term outcomes, where structured therapeutic patient education strategies integrated into MIC could play a decisive role [28].

Despite the important improvements in diabetes care goals, a proportion of T2DM patients did not reach the 3 goals targets set in this study. However, the study population had certain characteristics that prevented some participants from reaching the targets; for instance, the long-term diabetes diagnosis and potential risk for hypoglycemia/hypotension made them unsuitable for achieving ambitious care goals [29].. Compared with 1% of Mexican male patients and 12% of Americans [30], 25.9% of our study participants reached the three control goals.

In T2DM patients with poor self-care, perhaps conditioned to some degree by lack of knowledge, implementing diabetes self-management education (DSME) programs favors improving different interrelated activities performed by patients, including adhering to pharmacological and non-pharmacological recommendations [28]. The latter implies that multiple diabetes care goals beyond glycemic control, such as lipid and blood pressure control and weight reduction, are more likely to occur as a product of a stricter adherence and lifestyle modifications. An integrated team of health care professionals (HCPs) in a shared medical appointment model not only allows HCPs to personalize recommendations but also to strengthen health literacy and skills through a strategy of multiple same-day messages delivered by HCPs. This strategy facilitates creating shared decision plans and allows for follow-ups and feedback among T2DM patients and the team of HCPs. These care schemes have previously demonstrated to decrease HbA1c [31], LDL-C, and systolic blood pressure [32].

In terms of generalizability, our patients had characteristics particular to Latin-Americans with T2DM, including low-income and mixed genetic background. In terms of sex distribution, our study showed a higher number of women than men. This finding does not agree with the most recent National Health and Nutrition Survey – Mexico, which showed an equal sex distribution among T2DM in this age group [25]. This finding may suggest a potential population bias indicating that women tend to seek medical attention more often. Further research on how sex influences seeking medical attention in T2DM patients in Mexico City is warranted.

This study has several limitations. This is a before-and-after study, and individuals may not be comparable in their demographics and characteristics after 5 months. However, the strategies used in this study have been widely proven to be effective and are part of the current standards of care [33‐35]. Thus we reported real-world experiences in a public health care system providing services in a middle-income country. Also, the program results may not translate into long-term sustainability. Nevertheless, some studies have demonstrated that when implementing strategies similar to ours, long-term benefits are obtainable, including decreasing the risks of microvascular and macrovascular complications [36, 37] and all-cause mortality [38], independently of maintaining diabetes care goals, even if the immediate effect on diabetes care goals is minimal [39]. Moreover, patient empowerment and quality improvement interventions have demonstrated to be cost-effective in long-term empirical estimates [40, 41]. Additionally, despite the important improvements in diabetes care goals, a proportion of T2DM patients did not reach the target goals set in this study. However, the study population had certain characteristics that made these targets unreachable for some patients; for instance, the long-term diabetes diagnosis and the potential risk for hypoglycemia/hypotension made participants unsuitable for achieving ambitious target goals [29]. On the other hand, this study was conducted in Mexico City, a megapolis, and it is possible that our findings are not representative of other urban settings across the world. Nevertheless, the study highlights the need of considering SDH and ODH across populations that surely share similarities with our sample. Also, different factors regarding diabetes care goals achievement were not analyzed. However, diabetes care goals and diabetes-related outcomes have complex and multicausal origins, including biological, individual, and social factors. Our results are potentially affected by self-selection bias, particularly due to patients that did not agree to participate during the initial recruitment process or who were not interested in improving their diabetes management or outcomes. However, patients referred to the DIABEMPIC program came from primary level hospitals, and most of them agreed to participate and follow this program because of its free access and long-term health care benefits. 88% of the requested to participate accepted to follow the program and the rates of long-term participation in the program were 88%. Although we did not quantify the number of individuals included in this potential self-selection bias, our team reduced the possibilities of such bias: they made a remarkable effort to motivate and follow up all potential participants by making them and their family aware of the medical, economic, and social consequences of improper diabetes management. We also recognize that our study is limited by its design (uncontrolled before-and-after design). The limitations of our design include a lack of a control group and, therefore, a lack of randomization. However, we could not include a control group for ethical reasons (i.e., all the DIABEMPIC interventions have previously demonstrated a benefit in T2DM), and as the same individuals are followed after the intervention, this design is considered experimental [24]. Also, significant differences observed between time points in the outcomes may not have resulted from the intervention but might have been due to confounding. Even though the statistical analyses were adjusted for confounding by known variables, we recognize that it was not possible to record all characteristics that may influence the association between the intervention and outcome measures. Additionally, our study lacks a control group to determine the effect of the intervention, as well as the effect of SDH and ODH in this context. However, the components of the intervention program had an independent positive effect on diabetes clinical goals, including HbA1c. Therefore, to compare the intervention in a randomized study would fall within the ethical sphere of T2DM therapy.

Our findings support an encouraging proposal to address the overwhelming disease burden related to diabetes care, especially in populations with adverse SDH and ODH, where the implementation of quality improvement strategies probably lacks, but also where health equity in diabetes care can be achieved through such strategies. Long-term sustainability and cost-effective analysis are necessary, as well as quality of life and satisfaction of T2DM patients participating in these strategies. Diabetes care demands changing the role perspective of health care systems, HCPs, and T2DM patients to improve diabetes-related outcomes. MIC strategies could help to balance the co-responsibility, but the quality of care must be first assured in a favorable medical environment. Multicomponent initiatives with potential effectiveness in LMICs in T2DM must be replicated because they may have positive implications for population health and health care costs.