Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Assessing the influence of health systems on Type 2 Diabetes Mellitus awareness, treatment, adherence, and control: A systematic review

  • Suan Ee Ong ,

    Roles Conceptualization, Formal analysis, Methodology, Writing – original draft, Writing – review & editing

    suaneeong@u.nus.edu

    Affiliation Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore

  • Joel Jun Kai Koh,

    Roles Formal analysis, Methodology, Writing – original draft, Writing – review & editing

    Affiliation Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore

  • Sue-Anne Ee Shiow Toh,

    Roles Formal analysis, Writing – review & editing

    Affiliations Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore, Division of Endocrinology, Department of Medicine, National University Health System, Singapore, Singapore

  • Kee Seng Chia,

    Roles Writing – review & editing

    Affiliation Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore

  • Dina Balabanova,

    Roles Conceptualization, Formal analysis, Methodology, Writing – review & editing

    Affiliation London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Martin McKee,

    Roles Conceptualization, Formal analysis, Writing – review & editing

    Affiliation London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Pablo Perel,

    Roles Conceptualization, Formal analysis, Methodology, Supervision, Writing – review & editing

    Affiliations London School of Hygiene and Tropical Medicine, London, United Kingdom, World Heart Federation, Geneva, Switzerland

  • Helena Legido-Quigley

    Roles Conceptualization, Formal analysis, Methodology, Supervision, Writing – original draft, Writing – review & editing

    Affiliations Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore, London School of Hygiene and Tropical Medicine, London, United Kingdom

Assessing the influence of health systems on Type 2 Diabetes Mellitus awareness, treatment, adherence, and control: A systematic review

  • Suan Ee Ong, 
  • Joel Jun Kai Koh, 
  • Sue-Anne Ee Shiow Toh, 
  • Kee Seng Chia, 
  • Dina Balabanova, 
  • Martin McKee, 
  • Pablo Perel, 
  • Helena Legido-Quigley
PLOS
x

Abstract

Background

Type 2 Diabetes Mellitus (T2DM) is reported to affect one in 11 adults worldwide, with over 80% of T2DM patients residing in low-to-middle-income countries. Health systems play an integral role in responding to this increasing global prevalence, and are key to ensuring effective diabetes management. We conducted a systematic review to examine the health system-level factors influencing T2DM awareness, treatment, adherence, and control.

Methods and findings

A protocol for this study was published on the PROSPERO international prospective register of systematic reviews (PROSPERO 2016: CRD42016048185). Studies included in this review reported the effects of health systems factors, interventions, policies, or programmes on T2DM control, awareness, treatment, and adherence. The following databases were searched on 22 February 2017: Medline, Embase, Global health, LILACS, Africa-Wide, IMSEAR, IMEMR, and WPRIM. There were no restrictions on date, language, or study designs. Two reviewers independently screened studies for eligibility, extracted the data, and screened for risk of bias. Thereafter, we performed a narrative synthesis. A meta-analysis was not conducted due to methodological heterogeneity across different aspects of included studies. 93 studies were included for qualitative synthesis; 7 were conducted in LMICs. Through this review, we found two key health system barriers to effective T2DM care and management: financial constraints faced by the patient and limited access to health services and medication. We also found three health system factors that facilitate effective T2DM care and management: the use of innovative care models, increased pharmacist involvement in care delivery, and education programmes led by healthcare professionals.

Conclusions

This review points to the importance of reducing, or possibly eliminating, out-of-pocket costs for diabetes medication and self-monitoring supplies. It also points to the potential of adopting more innovative and integrated models of care, and the value of task-sharing of care with pharmacists. More studies which identify the effect of health system arrangements on various outcomes, particularly awareness, are needed.

Introduction

The 2015 International Diabetes Federation’s Diabetes Atlas [1] reported that 415 million people worldwide, or one in 11 adults, has diabetes, with most having Type 2 Diabetes Mellitus (T2DM) [1]. Although the incidence and prevalence of T2DM varies by geographical region, with over 80% of T2DM patients residing in low-to-middle-income countries, T2DM prevalence has increased worldwide since 1980 [2]. Health systems play a crucial role in the response to this rising burden, preventing premature death and disability and improving quality of life [2, 3]. Yet, while the management of diabetes has been the subject of many systematic reviews [412], these have focused on particular interventions, models of care, or information technology support systems. To our knowledge, no systematic review assembles the evidence appraising the impact of health systems on management of type 2 diabetes mellitus (T2DM). To address this gap, we systematically review the literature examining the health system-level factors influencing T2DM awareness, treatment, adherence, and control, and make recommendations for future research and policy considerations.

Methods

A protocol for this study was published on the PROSPERO international prospective register of systematic reviews (PROSPERO 2016: CRD42016048185). There are several ways in which the findings could be arranged but, as we were taking a health systems perspective, we used a conceptual framework developed by Balabanova and colleagues [13], which has been used to understand aspects of systems that hinder the effective management of non-communicable diseases [13, 14]. This framework identifies physical resources (e.g. healthcare facilities, pharmaceuticals, technologies), human resources (e.g. trained health workers), intellectual resources (e.g. clinical practice guidelines), and social resources (e.g. social capital, organisational measures to enhance collaboration) as necessary elements of a health system response to chronic disease challenges. The framework addresses inputs that underpin health system functioning in three key areas, namely service delivery, healthcare financing, and governance. These areas are recognised as critical elements of effective health system functioning by the World Health Organization [15]. In this review, following Gilson and colleagues, we define governance as: everyday actions and decisions that translate policy intentions into practice, which are “filtered through relationships, underpinned by values and norms, influenced by organisational structures and resources, and embedded in historical and socio-political contexts” [16] that reinforce or challenge institutional exclusion and power inequalities.

Additionally, this framework takes account of the critical role of context in influencing the health system. In this manuscript, “context” refers to the socio-demographic, economic, and cultural setting in which health systems are embedded and operate. This framework is guided by the understanding that health and healthcare systems are complex adaptive systems that are dynamic, evolving, have multiple constituent parts, and are often unpredictable, exhibiting path dependency and feedback loops [17, 18]. As such, the ability of a health system to produce good outcomes does not rest on the robustness of disparate constituent “blocks”, but on the integration and alignment of inputs and system functioning components [19]. This approach has several advantages. First, it ensures that all of the elements of the health system are considered explicitly. Second, by taking a health systems approach rather than, for example, a clinical approach based on detection, treatment and control, it is designed to facilitate identification of actionable points by health policymakers. Third, it identifies important gaps in the evidence that will be needed to develop a comprehensive health system response to diabetes, thereby contributing to prioritisation of research efforts. The corresponding disadvantage is that there will be some areas where there is little or no research. The conceptual framework is shown in Fig 1 below.

Inclusion and exclusion criteria

We included studies reporting the effects of macro and meso-level health system factors, interventions, policies, or programmes on T2DM control, awareness, treatment, and adherence. Box 1 outlines the definitions used and Box 2 details characteristics of included studies.

Box 1. Definitions of included T2DM outcomes

T2DM awareness: persons with clinically measured T2DM who have been diagnosed by a healthcare professional and are aware of their T2DM status

T2DM treatment: the use of at least one anti-diabetic medication in an individual with known T2DM

Anti-diabetic medication adherence: consistently taking antidiabetic medication as per regiment prescribed by a healthcare provider/professional

T2DM control: defined as the achievement of glycaemic control, blood pressure, and/or lipid control targets in individuals being managed for T2DM

Box 2. Characteristics of included studies

We included studies looking at any adult population, including general populations, populations receiving treatment, and populations of T2DM patients with related comorbidities, including hypertension and hyperlipidaemia. Included studies fell into two categories:

Studies undertaken at the macro-level of the health system: this includes, but is not confined to, national and international health policies, national healthcare financing structures, and national healthcare and health services delivery structures.

Studies undertaken at the meso-level of the health system: this includes, but is not restricted to, regional-level health systems/authorities, healthcare institutions (e.g. tertiary hospitals), and care organisations/networks (e.g. networks of primary care clinics)

Quantitative studies were included if they reported a measure of1 association between a health system element and at least one T2DM outcome. No date or language restrictions were applied. Translators were engaged to assist in determining the eligibility of non-English language literature. Translators helped with translation of titles, abstracts, and studies’ key findings. Studies evaluating interventions or programmes enacted at the micro-level (e.g. individual- or patient-level), such as those on the genetic profile of T2DM patients, were not included.

Search strategy

The search strategy drew on that used by Maimaris and colleagues [20] in their health systems and hypertension systematic review. Key words (MeSH terms) and free-text terms were identified for each domain of our health systems framework and combined with search terms for T2DM outcomes to generate search strategies for Medline, Embase, and Global Health. In addition, modified searches were performed on Latin American and Carribean Health Siences Literature (LILACS), Africa-Wide, Index Medicus for the South-east Asian Region (IMSEAR), Index Medicus for the Eastern Mediterranean Region (IMEMR), and Western Pacific Rim Region Index Medicus (WPRIM). All databases were searched from inception to 22 February 2017.

Study selection

Two reviewers independently screened search results by title and abstract for potential eligibility. Full-texts of potentially suitable articles were obtained and further screened by two reviewers. Disagreements were resolved by a third reviewer.

Data extraction for study setting, methodology, and findings

A data extraction form was created in Microsoft Excel. Two reviewers independently extracted data on design, setting/context, health system domain/s investigated, outcomes and relevant findings, and checked for disparities.

Risk of bias assessment

Two reviewers independently assessed included studies for risk of bias as low, medium, or high. For observational study designs (e.g. cross-sectional, case-control, cohort, pre-post, record/chart reviews) three domains were examined, as per Maimaris and colleagues in their systematic review [20]: selection bias, information bias (differential and non-differential misclassification), and confounding. Assessment of non-differential misclassification considered the reliability of the measure used to report T2DM outcomes. Studies assessed as having “low” or “high” risk of bias in most domains were classified as having low or high overall risk of bias respectively. Those where risk of bias was unclear in two domains, were classified as medium overall risk and those assessed to have unclear risk in three domains were classified as high overall risk.

The Cochrane risk of bias tool was used to assess randomised controlled trials, cluster randomised trials, and non-RCT, non-observational studies (e.g. trials that are not randomised or do not have a control group). Studies assessed as having low risk of bias across most domains were classified as low overall risk of bias. If risk was unclear in two to three domains and most domains were not classified as “high” or “low” risk of bias, the study was classified as medium overall risk of bias. Studies assessed to have unclear risk of bias in four domains were classified as having high overall risk of bias.

For quality assessment of qualitative studies, we used an adapted version of a checklist previously used in a series of mixed-methods systematic reviews [21, 22], comprising ten core criteria. Studies with a score of eight to ten were classified as having an overall low risk of bias, four to seven as overall medium risk of bias, and zero to three as overall high risk of bias.

Assessment of context and complexity considerations

We assessed the extent to which included studies consider context and complexity, in respect of sociodemographic, political, economic, and/or cultural issues, as well as dynamic relationships between different health systems domains. We also explored how health systems interacted with contextual factors.

Data analysis and synthesis

A narrative synthesis was performed. We categorised studies by health system domain and study setting, recognising that some investigated multiple domains simultaneously. Randomised controlled trials (RCTs) were considered the strongest design to establish causality, followed by cohort and case-control studies. Cross-sectional and ecological studies were not considered adequate to establish causality. We did not conduct a meta-analysis due to heterogeneity across study designs, populations, comparisons, analytical strategies, and outcomes [20].

Results

We describe the screening process using an adapted Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart [23], shown in Fig 2 below.

Database searching identified 6,975 records, with 5,620 remaining after duplicate removal. After screening of titles and abstracts, 175 full-text articles were retrieved. 93 were included in the final qualitative synthesis. Of these 84 were quantitative. Of these, 21 were randomised controlled trials; one was a cluster randomised controlled trial; three were cluster randomised pragmatic trials; three were trials (i.e. trial designs with no mention of randomisation); 15 were cohort studies; one was a case-control study; 19 were cross-sectional studies; 14 were pre-post studies; six were record/chart reviews; and one was a time-series analysis. Of the remaining nine studies, six were qualitative and three used mixed methods. 77 (83%) of included studies were carried out in World Bank-classified high-income countries, nine in upper-middle income countries, and seven in lower-middle income countries.

Geographical distribution of included studies

As shown in Fig 3, most studies took place in North America (n = 53) and Europe (n = 16), with fewer in Asia (n = 10), Africa (n = 6), South America (n = 1), the Middle East (n = 3), and Australia (n = 4). Notably, all studies of healthcare financing were conducted in North America (n = 14).

Risk of bias assessment

We conducted risk of bias assessment for all 93 articles. 28 studies had their risk of bias assessed using the Cochrane risk of bias tool. Of these, 22 were randomised controlled trials. 12 had high risk of bias [2435], eight had medium risk of bias [3643] and one had low risk of bias [44]. One cluster randomised controlled trial was assessed to have low risk of bias [45]. Three were cluster randomised pragmatic trials; two had high risk of bias [46, 47] and one had low risk of bias [48]. Three studies were trials: two had high risk of bias [49, 50] and one had medium risk of bias [51].

Among the 56 studies assessed as observational study designs, 19 were cross-sectional, of which seven had high risk of bias [5258], four had a moderate risk of bias [5962] and eight had low risk of bias [6370]. 15 were cohort studies; one study had a high risk of bias [71], five had a medium risk of bias [7277], and eight had low risk of bias [7885]. One study used a case-control study design and had a moderate risk of bias [86]. 14 were pre-post studies; one had a high risk of bias [87], 11 had medium risk of bias [8898], and two had low risk of bias [99, 100]. Six studies were record/chart reviews: two were assessed to have low risk of bias [101, 102] and four had medium risk of bias [103106]. The one time-series study in this review had a medium risk of bias [107].

Six qualitative studies were assessed for risk of bias using a tool adapted from the Consolidated Criteria for Reporting Qualitative Studies (COREQ) checklist. One had a moderate risk of bias [108], and five had low risk of bias [109113].

Risk of bias in the three mixed-methods studies was assessed separately for the quantitative and qualitative components. All three studies had quantitative components with a high risk of bias. One study had a moderate risk of bias for the qualitative component [114] and two had a low risk of bias for the qualitative component [115, 116].

Context

32 of 93 included studies gave no detailed information about the socio-demographic, political, or economic context in which the study was conducted [26, 2830, 34, 36, 37, 39, 40, 44, 45, 47, 52, 53, 55, 57, 64, 65, 68, 70, 72, 73, 84, 87, 89, 90, 95, 97, 101, 103, 105, 110]. 61 included studies that provided contextual information on various levels.

Three studies described the regional context, such as Sub-Saharan Africa and South America [49, 83, 99]. Such studies tended to consider context within the narrative of diabetes control in their region. 32 studies described the national context in which the study took place [25, 27, 31, 33, 42, 43, 48, 50, 51, 54, 5962, 66, 67, 74, 8082, 85, 86, 91, 92, 96, 98, 100, 104, 111, 112, 115, 116]. This typically involved descriptive statistics to indicate the magnitude and urgency of diabetes as a national challenge, highlighting incidence or prevalence rates and cost burden. These descriptions ranged from brief summaries to comprehensive elaborations. 16 studies described the health system context in which the study took place [32, 35, 46, 58, 69, 75, 78, 88, 93, 94, 102, 106110]. The contexts described ranged from broad and general to in-depth and extensive.

10 studies considered the context of the population studied or the specific intervention. Examples included the role of healthcare professionals involved in the intervention [38, 77], descriptions of demographic context (e.g. low-income, indigenous, Hispanic) in which an intervention took place [24, 41, 71, 76, 79, 113, 114], and existing structures in which interventions occurred (e.g. financing of prescription medications in the Veterans Affairs healthcare system) [56]. Table 1 provides examples of context considerations in included studies.

Effect of health system inputs on diabetes outcomes

37 studies explored the impact of health system inputs on diabetes outcomes. The analysis explores studies that had a focus on one type of resource, and complex interventions involving studies with more than one type of resource or system building block. Table 2 summarises the findings of included studies exploring the associations between health systems inputs and diabetes outcomes.

thumbnail
Table 2. Summary of findings of studies examining the associations between health systems inputs and T2DM outcomes.

https://doi.org/10.1371/journal.pone.0195086.t002

Physical resources.

One cross-sectional study from the United States found an association between driving distance from primary care facility and likelihood of insulin use (odds ratio (OR) for using insulin associated with each kilometre of driving distance, 0.97, 95% confidence interval (CI) 0.95, 0.99; p = 0.01) [68]. Living within 10km of a primary care facility was associated with increased likelihood of insulin use (OR 2.29, 95%CI 1.35–3.88; p = 0.02).

Human resources.

36 studies examined the effect of human resource inputs on diabetes outcomes.

Pharmacists: 17 quantitative studies explored the impact of pharmacists on diabetes outcomes. All but one [88] took place in high-income countries. 11 found positive effects of pharmacist care on diabetes control and adherence outcomes.

Seven studies from high-income countries reported positive impacts of pharmacists administering patient care. Two trials, both conducted in the United States, obtained positive results. One found a significantly greater absolute percent decrease in HbA1c from baseline among those seeing a pharmacist [30]. The other, in university owned neighbourhood clinics, reported that the intervention group achieved reductions in HbA1c with fewer physician visits compared to patients receiving usual care [33].

A cohort study in the United States reported a lower mean HbA1c among those in an outpatient programme involving face-to-face pharmacist consults (p = 0.024), and significantly reduced from baseline [76]. Although there was no significant between-group difference in mean medication possession ratio (MPR), intervention patients saw an increase in MPR from baseline (p = 0.024). Patients receiving the intervention were less likely to discontinue diabetes medications (p<0.001) and more likely to have their medication prescription 30 days after the end of their supply of the last prescription following their first consultation date (p< 0.001).

An Australian randomised controlled trial found that a pharmacist-delivered community-based care and management programme was associated with a mean decrease in blood glucose levels over six months [32]. Improvements in HbA1c were greater in the intervention group (-0.97% (95% CI: -0.8, -1.14, p<0.01) compared to controls (-0.27% (95% CI: -0.15, -0.39, p<0.01)).

A pre-post study from the United States found that an intervention with pharmacists adjusting medications, evaluating therapeutic needs, and developing care plans recorded significantly lower mean HbA1c readings compared with the control group after one and two years. The two-year average decrease in HbA1c was greater for the intervention group compared to the control (p = 0.009) [100].

Another pre-post study from the United States found that involving clinical pharmacists in direct patient care of insulin-dependent patients in primary care led to a significant decrease in mean HbA1c post-intervention [94]. A Maltese pre-post study found that an education-focused pharmacist intervention was associated with a smaller proportion of patients “rarely” missing a dose of medication and a larger proportion of patients reporting “never missing a dose” of medication post-intervention [98].

Four studies reported positive impacts of pharmacist-led patient management. A randomised controlled trial in the United States found that pharmacist-led shared medical appointments was associated with significant reductions in HbA1C (-0.41; 95% CI -0.74 to 0.07) and significantly higher odds of attaining HbA1C goals (adjusted OR 2.73; 95% CI, 1.03 to 7.16) compared to usual care [26]. A pre-post study in the United States found that a pharmacist-led, patient-centred pharmacotherapy management programme was associated with significantly reduced HbA1c and fasting plasma glucose parameters for patients with diabetes who had complex disease and medication burdens at six and 12 months when compared to baseline [91]. A Nigerian pre-post study found reductions in mean HbA1c and fasting blood sugar in those receiving monthly follow up pharmacists over three months at a primary healthcare facility [88]. A pre-post study in the United States found that having a pharmacist practice in a patient-centred medical home (a team-based model of care led by a personal physician who provides continuous and coordinated care throughout a patient’s lifetime to 15 outcomes) was associated with significant decreases in patients’ mean HbA1c from baseline at one- and two-year time points [89].

Four studies, all from North America, found mixed effects of pharmacist involvement in service delivery on diabetes control and adherence outcomes. A Canadian randomised controlled trial [42] found a reduction in systolic blood pressure (SBP) of 7.4mmHg (95%CI 4.6–10.2; p<0.001) over one year in patients managed by a team to which a pharmacist had been added, with a between-group difference in SBP of 4.9mmHg, 95%CI 1.0–8.7; p = 0.01. However, there were no significant changes in glycaemic control and lipid parameters. A pragmatic cluster randomised trial in the United States found that a clinical pharmacist-led outreach programme (working with patients in person or over the phone, motivational interviewing) had only short-term positive effects for primary care patients [46]. Immediately post-intervention, the mean SBP of the pharmacist-led group decreased compared to the control group (-2.4mmHg, 95%CI 1.5–3.4, p<0.001). However, the control group achieved similar results six months post-intervention. A United States cohort study found that a face-to-face community pharmacist intervention focused on counselling and education for Hispanic patients with diabetes led to reductions in fasting plasma glucose (p = 0.019), SBP (p = 0.003), and triglycerides (p = 0.003) from baseline to 12 months, but not in mean HbA1c. Subgroup analyses of patients with poorly controlled diabetes at baseline showed significant reductions in mean HbA1c, SBP, diastolic blood pressure (DBP), and lipids [71]. A record/chart review in the United States found that diabetes management led by clinical pharmacists [102] was associated with a reduction in mean HbA1c and an increase in patients achieving HbA1c <7% over four years. However, no significant improvements were found in SBP, DBP, lipid measures, or medication adherence.

Two studies found no impact of pharmacists on outcomes. A randomised controlled trial in the United States exploring the impact of a pharmacist intervention (including meetings and phone-calls to initiate care plans) on poorly-controlled patients with diabetes managed in primary care found no difference in mean HbA1c or self-reported medication adherence [33]. A cohort study in the United States found no significant effect on medication adherence associated with the presence of a pharmacist in Veterans Affairs clinics [77].

Nurses: Eight studies, all quantitative, looked at the impact of nurses on diabetes control and adherence outcomes. All but one [99] took place in high-income countries. Three studies found positive impacts on control of service delivery by nurses. A Dutch randomised controlled trial explored the impact of transferring routine aspects of diabetes care in hospital outpatient clinics to diabetes specialist nurses [29]. After one year, significantly more patients receiving care from nurses achieved HbA1c <7% compared to the control group. A Cameroonian pre-post study [99] found a significant reduction in mean fasting blood glucose (FBG) in non-insulin-dependent patients’ following a nurse care empowerment scheme. A Danish cross-sectional study [67] found that the proportion of patients with HbA1c ≥8% significantly decreased in general practices with well-implemented nurse-led diabetes consultations, compared to practices without.

One record/chart review in the United States reported mixed results. It evaluated the impact of incorporating nurse practitioners into collaborative practices with primary care clinicians [106] and found that, post-intervention, 50% of patients achieved HbA1c <8% compared to 0% of patients pre-intervention (p<0.001). There were no significant changes in the proportion of patients achieving BP and cholesterol targets.

Four studies, all trials, reported no significant effects of service delivery by nurses on diabetes control. Two were conducted in the Netherlands and examined the impact of diabetes management by nurses [27] and nurse-driven, protocol-based correctional therapy on patients [50]. Two were conducted in the United States and explored the effect of nurse-led behavioural management [37] and empowerment of nurse practitioners to provide comprehensive patient care [31].

Physicians: Six studies, all quantitative and conducted in high-income countries, investigated the impact of physicians on diabetes outcomes. A cohort study in the United States found that 10% increased frequency of therapeutic intensification (i.e. increasing the dosage or amount of hypoglycemic medication a patient takes) by a physician was associated with a 0.15% lower level of HbA1c (p<0.0001) among patients in an urban health system. A single episode of therapeutic intensification was associated with an average 0.7% reduction in HbA1c (p<0.001) [85].

A record/chart review in the United States reported that newly-diagnosed patients with diabetes at a family medicine clinic of a university hospital with a regular physician reported lower mean HbA1c values than patients without a regular physician [103]. A Dutch record/chart review found that symptom-catalysed (encouraged by the onset of symptoms in the patient), opportunistic, or patient-requested general practitoner (GP) screening activity was significantly related to the presence of a diabetes diagnosis [101].

A review of United States claims records found a positive association between physicians’ certification by a national-level quality assurance organisation and patient adherence [105]. Patients managed by certified physicians were more likely to receive oral anti-hyperglycaemic drug prescriptions (mean prescriptions per patient per month 0.49 vs. 0.46, p = 0.02).

An Australian cross-sectional study compared the impact of vocationally-registered vs. non-vocationally-registered general practitioners (GPs) on diabetes control [59]. Vocational registration entails the enrolment of GPs as part of improving professional standards, rewarding high-quality practice, and enabling GPs’ access to higher rebates in the publicly-funded universal healthcare system. It found no difference in mean HbA1c. A United States cross-sectional study [70] found an association between patient-physician gender concordance and diabetes control, with female patients of female physicians most likely to have HbA1c <8%. However, this was not due to differences in medication adherence.

Community health workers: Three studies focused on the impact of community health workers (CHWs). Two reported positive impacts on diabetes outcomes. A ixed-methods study in the United States found that a culturally-relevant diabetes self-management education programme led by trained, bilingual CHW to improved mean HbA1c and SBP among uninsured and underserved Hispanic patients [114]. An Australian trial found that poorly-controlled indigenous patients managed in primary care and receiving a clinical, management, education, and social and family support-focused CHW intervention by a trained indigenous CHW resident in the community had improved mean HbA1c levels compared to controls [41]. Meanwhile, a randomised trial of Hispanic/Latino patients with diabetes in the United States managed at family health centres found no differences in control and adherence between patients receiving care delivered by full-time, trained bilingual CHWs who had T2DM/had experienced it via a family member or friend vs. case management and standard care [24].

Peers: Two trials investigated the impact of peers on diabetes outcomes. Results were mixed. A randomised controlled trial in the United States of veterans with poor glycaemic control receiving usual nurse care vs. reciprocal peer support (i.e. an intervention that encouraged patients with diabetes to receive and provide support to each other) [44] found that peer support significantly impro17ved mean HbA1c levels while in those receiving nurse-led care mean HbA1c levels increased. Among patients with baseline HbA1c <8.0%, patients receiving peer support (-0.88%) demonstrated greater improvement in mean HbA1c (-0.07%) compared to those receiving nurse care. Meanwhile, an Irish cluster randomised controlled trial compared outcomes of patients receiving standardised diabetes care vs. a two-year peer-support intervention [45]. At two-year follow-up, there were no significant differences in HbA1c, SBP, and total cholesterol.

Intellectual resources.

No included studies evaluated the relationship between intellectual resources, such as use of guidelines, and diabetes outcomes.

Social resources.

No included studies evaluated the effects of social resources on diabetes outcomes.

Health systems financing

21 studies examined the effect of health systems financing on diabetes outcomes. 15 were quantitative and conducted in high-income countries: 13 in the United States and two in Canada. Six were qualitative, and all but one [111] took place in lower and middle-income countries. The high-income country study [111] was the only study not reporting cost or financial difficulty as a barrier to diabetes control, treatment, or adherence. Table 3 summarises the findings of included studies exploring the associations between healthcare financing and diabetes outcomes.

thumbnail
Table 3. Summary of findings of studies examining the associations between healthcare financing and T2DM outcomes.

https://doi.org/10.1371/journal.pone.0195086.t003

Cost-sharing and outcomes.

Five studies in the United States examined the relationship between cost-sharing (i.e. when patients pay out-of-pocket for a portion of healthcare costs not covered by health insurance) on outcomes. All studies found that adherence and/or control measures decreased as cost-sharing increased. These findings were consistent across different types of cost-sharing and health schemes [60, 65, 74, 78, 80].

Insurance status and outcomes.

Four studies looked at the impact of insurance status on outcomes. All were from the United States and cross-sectional. One found that publicly-insured patients had significantly lower mean HbA1c values (p<0.001) and better control (HbA1c ≤7%, p<0.001 than those with private insurance [66]. A study of elderly patients found that privately insured patients were almost twice as likely to report diabetes medication underuse compared to patients in the Veterans Affairs system (p<0.0001) [56]. Another study using national data [62] found the highest diabetes control among commercially-insured patients, compared to those covered by Medicare or Medicaid, but adherence was higher among Medicare beneficiaries. Another explored the impact of socioeconomic factors on control among newly-diagnosed patients with diabetes participating in a community-wide screening programme [53] and found that lack of insurance coverage was predictive of patients failing to seek medical care.

Extent of healthcare insurance and outcomes.

Six studies looked at the impact of different levels of healthcare insurance coverage (i.e. the extent to which different care services, treatment options, medications, and/or self-monitoring and testing equipment are covered under a healthcare insurance pan) on outcomes. Four found that broader coverage was associated with better outcomes. Two, a time-series study among patients with a health maintenance organisation in the United States [107] and a Canadian cross-sectional study of patients receiving care at community pharmacies [52], found that the provision of free monitors/testing supplies (e.g. blood glucose monitors, insurance coverage for supplies e.g. glucometer strips) improved control.

Two studies, both from the United States, found that increased coverage of drugs was associated with improved adherence. A cohort study examined the relationship between Medicare Part D benefit coverage (a federal government programme to subsidise the cost of prescription drugs and drug insurance premiums for Medicare beneficiaries) [75], finding that beneficiaries without coverage (OR 0.617, P<0.0001, 95% CI 0.523, 0.728) or generic coverage only (OR 0.702, P<0.001,95% CI 0.604, 0.816) were less likely to be adherent than those with full coverage of generic and branded drugs. A pre-post study found that uninsured outpatients participating in a pharmacy-managed medication programme providing free medication [92] had significant reductions in mean HbA1c (p = 0.002) and total cholesterol (p = 0.001), and an increase in proportion of patients achieving HbA1c <7%. However, a Canadian randomised controlled trial of free (versus out of pocket payment) self-monitoring supplies [39], found no difference in six-month HbA1c between intervention and control groups. Only one cohort study examined the relationship between type of healthcare financing plan and adherence among Medicaid patients from across the country [82]. It found that patients in capitated plans had 5% lower mean oral antidiabetic adherence than those in fee-for-service plans (p<0.05).

Impact of financial factors on outcomes.

Six studies, all qualitative, reported on the impact of financial factors (e.g. cost of services, medication, lifestyle management, and the ability of persons with diabetes to pay for them) on outcomes. A Tunisian study using interviews with patients and healthcare professionals in primary care settings [109] and an Indian study using interviews with patients who have diabetes living in urban slums [110] both found financial factors to be a key barrier to access to medication, affecting adherence to diet, medication, blood tests, and referrals. A Bangladeshi study of interviews with patients who have diabetes managed at various care facilities [108] reported that availability and cost of services impeded access to appropriate diagnosis and subsequent treatment.

A Kuwaiti study using interviews with patients who have diabetes and are on oral medication managed in general practice or hospitals [112] reported unavailability of medications, difficulties accessing physicians and medications, inequalities in care provision and medication supply at different healthcare facilities, and lack of trust in the government healthcare system as barriers to adherence. A South African study interviewed low-income female patients with diabetes [113] and found that patients’ adherence to medication was affected by structural factors in the health system, including overcrowded clinics and poor access to medicines. However, a series of study of focus groups conducted in the United States with economically-disadvantaged urban-dwelling African-American patients with diabetes [111] found that the main contributors to lack of medication adherence were denial of consequences and a lack of understanding of the disease, and not cost or financial concerns.

Service delivery

26 studies investigated the relationship between health service delivery and diabetes outcomes. Table 4 summarises the findings of studies examining associations between service delivery and diabetes outcomes.

thumbnail
Table 4. Summary of findings of studies examining the associations between service delivery and T2DM outcomes.

https://doi.org/10.1371/journal.pone.0195086.t004

Innovative/integrated models of care.

14 studies examined the impact of innovative/integrated care delivery models on outcomes. In this review, we define innovative/integrated care delivery models as multifaceted care models that bring together different components of services towards improving outcomes.

Five studies took place in Europe: one each in Spain [84], the Netherlands [73], Denmark [47], the United Kingdom [79], and Italy [97]. Three studies were set in Asia: two in China [25, 43] and one in Thailand [86]. Three studies took place in the United States [87, 104, 116], and one each was in Australia [51], the Middle East [93], and Central America [83].

Despite variation in types of innovation/integration of care employed, implementation sites, and country settings, 11 studies found positive associations between innovative/integrated models of care delivery on diabetes control and adherence outcomes. A Chinese trial among older patients with diabetes receiving hospital-based specialist care [25] found improved mean HbA1c in the group receiving integrated care compared with the traditional model. An Australian trial [51] found that mean HbA1c at 12 months significantly decreased in a group receiving integrated community care while there was no significant change in a control group.

Three cohort studies found improved outcomes. A Dutch study [73] found that mean HbA1c and the proportion of patients with poor control both fell significantly in a group receiving structured care (i.e. the implementation and practice of care processes as per clinical practice guidelines) from general practitioners but rose in the control group during 2 years of follow-up. Good control (HbA1c <7%) was achieved in 54.3% of those receiving structured care compared to 44.1% of controls (p = 0.013). A Mexican study of group management found that mean fasting plasma glucose and BP were significantly lower at 15 months’ follow-up than in controls receiving usual treatment [83]. A Spanish evaluation of a multifactorial intervention in primary care centres [84] found highly significant improvements in a wide range of biochemical parameters after one year.

A Thai case-control study [86] found significantly lower mean HbA1c and a significantly higher proportion of patients with HbA1c <7% in those managed in a community clinic emphasising promoting continuity of care compared to those receiving standard care in an outpatient hospital setting. A cross-sectional study in the United Kingdom found that an integrated "care package" [79] was associated with lower HbA1c values and increased probability of meeting cholesterol (≤4mmol/l) and BP (≤140/80mmHg) targets compared with controls receiving usual treatment.

Three other quantitative studies found positive results: a pre-post study in the United States assessing the impact of using the Asheville care management model [87], a record/chart review also in the United States looking at effects of a diabetes management programme in a university health system [104], and an Italian pre-post study assessing the impact of a structured education-based model at a diabetes clinic [97]. All three found improvements in achievement of control outcomes (e.g. proportion of patients achieving HbA1c, SBP, DBP, and cholesterol goals). A mixed-methods study in the United States evaluated a ‘Patient Navigator Model’ [116], finding lower mean HbA1c post-intervention (7.8% vs 7.2%, p = 0.01).

Two trials obtained mixed results. A Chinese trial of a case management, behaviour change-focused, protocol-driven model of care in an outpatient hospital setting [43], found that mean HbA1c was significantly reduced in the case management group at six months. A Danish trial of a structured personal care model [47] found a significant benefit only among female patients: the median HbA1c level was 8.4% in women receiving structured care vs. 9.2% in women receiving usual care. A pre-post study in the United Arab Emirates found no effect of an integrated care model on outcomes in primary care facilities [93].

Coaching and education.

Nine studies, all quantitative, focused on the impact of delivering services including coaching and education. Four found positive effects on control and adherence.

An Israeli trial compared the impact of standard care with a participative programme including education and lifestyle modification among patients with diabetes, hypertension, and hyperlipidaemia [28]. Over a mean follow-up of 7.7 years, mean HbA1c, SBP, DBP and LDL-C values were significantly lower in the intervention group. A US trial found that nurse and dietitian-administered diabetes education and health behaviour classes delivered in primary care were associated with significant reductions in mean FBG, mean HbA1c, mean LDL-C, and mean total cholesterol [34]. A Nigerian trial examining the impact of prolonging physician-patient interaction and incorporating health education forums for patients with diabetes found a significant difference in the percentage of patients who had good fasting glycaemic control and 2-hour post-prandial glycaemic control in the intervention compared to the control group [49].

A pre-post study in the United States found that an outpatient diabetes education programme delivered by certified diabetes educators, who are either registered nurses or dietitians, was associated with decreased mean HbA1c (p = 0.001) and increased medication adherence for antihypertensive agents, aspirin, injectable insulin, and insulin sensitisers (p = 0.001). Among patients with uncontrolled diabetes (HbA1c >7.0%), there was a significant decrease in mean HbA1c between baseline and follow-up [90].

Two studies obtained mixed results. A randomised controlled trial conducted in the Netherlands found that providing counselling to patients managed in general practice had no significant effect on mean HbA1c after a year [38]. However, among those with FBG >10mmol/l, the mean HbA1c of intervention group patients was lower than that among control group patients (p = 0.001). A British randomised controlled trial examining the impact of a healthcare professional-delivered structured group education programme delivered in the community found no statistically significant differences in mean HbA1c of the structured group education arm compared to the control group [36]. However, the intervention group showed a reduction in triglyceride levels at eight months (intervention -0.57mmol/l (-0.71–-0.42) vs. control -0.34 mmol/l (-0.53–-0.15), p = 0.008).

Three studies–a Chinese randomised pragmatic trial of telephone-based motivational interviewing (MI) and health coaching [48], a South Korean pre-post study of individually-tailored diabetes education by visiting nurses on low-income patients [96], and an American randomised controlled trial of provision of adherence advice and/or MI [40] found no significant impact of coaching and education-based interventions on control and adherence.

Healthcare type/setting.

Two studies, both quantitative, looked at the impact of healthcare type/setting on diabetes outcomes. A Brazilian cohort study compared those attending a public and private clinic [81]. While both groups showed improvements across all clinical parameters, patients receiving public healthcare had significantly higher mean HbA1c and mean cholesterol than private care patients. A Taiwanese cross-sectional study explored differences in control outcomes among patients receiving care in primary vs. secondary/tertiary healthcare settings [61] and found that patients in primary care had higher mean HbA1c values but better BP control than patients in secondary/tertiary care.

Access and use.

A cross-sectional study in the United States examined the impact of access to healthcare on diabetes control among patients receiving care from a large two-county public health system delivering services to a vulnerable, high-risk minority ethnic group population. This study looked at different facets of access including the availability of insurance coverage, experience of seeking care, and the ability to access medication. Having trouble getting care was associated with a 0.57% increase in HbA1c (p = 0.04), use of an acute care facility was associated with 0.49% higher HbA1c (p = 0.047) and having gone nowhere for care was associated with 1.08% higher HbA1c compared to going to a doctor's office. Lack of insurance was not found to be associated with levels of HbA1c [57].

Governance

Two included studies evaluated the effects of health system factors relating to governance challenges. Table 5 summarises the findings of studies examining associations between governance and diabetes outcomes.

thumbnail
Table 5. Summary of findings of studies examining the associations between governance and T2DM outcomes.

https://doi.org/10.1371/journal.pone.0195086.t005

Two cross-sectional quantitative studies looked at the impact of patient-physician relationships on diabetes outcomes. Both studies found that the nature of patients’ relationships with physicians impacted control and/or adherence. A Danish study found that patients whom GPs classified as not knowing well had relatively higher mean HbA1c and fasting plasma glucose levels compared to those of patients classified as known “well” or “fairly well” [64]. A study in th United States found that among patients reporting high levels of physician trust, rates of cost-related medication underuse significantly increased from 4% among patients with low monthly out-of-pocket costs (<US$51) to 11% among patients with high monthly out-of-pocket costs (>US$100) [55].

Complex interventions: Studies with more than one building block

Seven studies evaluated outcomes incorporating components from multiple health systems domains. Three took place in high-income countries [54, 72, 95], five were quantitative [54, 63, 69, 72, 95], and one was mixed-methods [115]. Three studies looked at the combined impact of more than two health systems components. Table 6 summarises the findings of studies examining associations between complex interventions and diabetes outcomes.

thumbnail
Table 6. Summary of findings of studies examining the associations between studies with more than one health systems building block and T2DM outcomes.

https://doi.org/10.1371/journal.pone.0195086.t006

A South African mixed-methods study looked at the impact of service delivery, intellectual inputs, and governance, examining the effect of a programme combining the Chronic Care Model with primary care nurse support, medication scale-up, and improved access to specialist care [115]. It reported improved disease awareness through early detection and detecting patients with advanced disease, and improved treatment through referral of high-risk, poorly controlled patients. A German prospective survey found that systems-level changes, including specialist physicians, structured teaching and training programmes, postgraduate training courses for physicians and staff, and increased access to self-monitoring equipment were associated with improvement in mean HbA1c and percentage of patients with mean HbA1c <7.2% between 1989 to 2000 [72]. A Mexican cross-sectional study assessing the impacts of healthcare financing and physical and human resource inputs [58] found that uninsured patients were more likely to have mean HbA1c >12.0% than insured patients, that municipalities with more health units in relation to population had more patients with poor HbA1c (OR: 3.17), and insured patients and those living in areas with more nurses per 1,000 population had a greater likelihood of not having poor HbA1c (OR: 4.59).

Remaining studies looked at the combined impact of two health systems components. Two cross-sectional studies, one each from the United States [54] and the Philippines [69], looked at the combined impact of service delivery and intellectual inputs, with a focus on the concepts/principles of the Chronic Care Model. Both studies found positive relationships between service delivery founded in the Chronic Care Model and diabetes control (e.g. mean HbA1c) and awareness (e.g. odds of being tested for HbA1c, FBG, lipids) outcomes.

A Malaysian cross-sectional study examining the impacts of service delivery and human resource inputs on diabetes control [63] found that patients at a hospital with specialists had significantly lower mean HbA1c, LDL-C, and higher mean HDL-C than patients at a health clinic with family physicians, a health clinic without a doctor, and a hospital without specialists. Patients at a health clinic with family physician were most likely to achieve HbA1c ≤ 6.5% (aOR 1.2) and BP targets (aOR 1.4). Patients at a hospital without specialists were 3.4 times more likely to not achieve LDL-C targets. A pre-post study in the United States examining the impacts of service delivery and healthcare financing on outcomes [95] found a performance-based provider compensation programme significantly increased the probability of disadvantaged, underserved patients receiving two HbA1c tests by 15.67%.

Health systems complexity considerations

67 of the 93 studies did not address interdependence and linkages between health system domains [2530, 3340, 4246, 4852, 56, 59, 61, 62, 6568, 7278, 8087, 8995, 97, 98, 103107, 115, 116]. Of the studies which did, three considered linkages between human inputs and service delivery [69, 88, 99]. These studies considered the need for sufficiently trained healthcare professionals beyond simply delivery of healthcare services on outcomes.

Five studies considered interdependencies between healthcare financing and service delivery [32, 55, 100, 102, 114]. They pointed to high costs of associated services required for diabetes care and control, particularly with regard to the cost of prescription medication. One study pointed to a lack of understanding among clinicians, arguing that “physicians’ role in influencing patients’ response to medication costs [are] not well understood” [55].

Four studies considered the linkages between multiple, dynamic health system domains [108110, 112]. These tended to study health system barriers and facilitators of diabetes care using qualitative methodologies. For example, a Tunisian study invited patients and healthcare professionals to discuss the barriers they faced in receiving and providing care. The findings were positioned in relation to several health system domains, namely inputs (i.e. human, physical, intellectual resources), service delivery, and healthcare financing [109].

14 studies considered the interactions of other factors not usually covered within the health system building blocks, but still linked to the health system. Five considered patients’ socioeconomic status [53, 57, 64, 101, 111], of which two discussed issues facing low-income patients [58, 96]. Five studies considered ethnic factors [24, 41, 60, 71, 113]. Two studies considered the role of gender [47, 70]. These studies explored how patients’ demographic characteristics might interact with different health system domains and affect outcomes.

Discussion

This systematic review sought to assess the influence of health systems-level factors on T2DM awareness, treatment, adherence, and control. Despite the limited scope and variable quality of included articles, we identified several key health system facilitators and barriers to diabetes control, treatment, adherence, and awareness outcomes. Several studies also reported on health system factors which were neither facilitators nor barriers, showing no statistically significant impact on outcomes.

We confirmed the importance of two important health system barriers to effective diabetes care. The first was the presence of financial constraints faced by the patient. These were either self-reported or implied by the presence of co-payments for medication and were found to be a barrier to control and/or adherence outcomes. We also showed that reducing out-of-pocket payments could improve diabetes outcomes. Second, lack of access to health services and medication was a barrier to achieving good diabetes outcomes, with the evidence primarily from qualitative studies.

We also found three health system facilitators. First, integrated, innovative care models were positively associated with improved diabetes outcomes. These results are consistent with studies of integrated care programmes for other chronic diseases [117, 118]. Despite the heterogeneity of integrated care models and the national settings where they were implemented, many studies that yielded positive results had certain commonalities, namely patient education/empowerment, care continuity, data management, task-sharing, and multidisciplinary/team-based care. Yet it remains unclear what is the right mix of integrated care components and whether some are more important than others. Also, as most of these studies were in high-income countries, their applicability to low- and middle-income settings is unclear. It is also important to note the lack of consensus on definitions of “innovative” and “integrated care” [119, 120] and the components/elements that constitute an “integrated” care model [121, 122] in the health systems research community. Second, there is evidence to support pharmacist involvement. While this is consistent with findings in other chronic disease outcomes, as well as several meta-analyses [123, 124], we cannot rule out the potential of publication or reporting bias informing our findings. In this review, we surprisingly found fewer papers on nursing that we did pharmacists. This may be because nursing models are better established and therefore less novel, with fewer publications relevant to the aims of this review–but not necessarily less effective or impactful than pharmacist care models. Third, education programmes led by health professionals showed mostly positive effects on control.

Several other health systems facilitators were identified, namely peer support, positive patient-physician relationships, and multi-faceted interventions, but there is rather less evidence to draw on.

Study strengths and limitations

This review has several strengths. It includes a wide range of measures of control, treatment, awareness, and adherence. Its conceptual framework, building on previous work by Maimaris et al [20] and Balabanova et al [13], enabled studies to be linked to different health systems domains. Our systems approach helps us map the global landscape of health systems-related research on diabetes outcomes, identifying geographical and topic gaps in research and variation in types and rigour of studies conducted, thereby informing decisions on needs for future research.

Another strength is the inclusive view of governance that encompasses the relationships and everyday practices of actors in the health system. By regarding governance as practices that are driven by macro- or meso-level decision-making, but operationalised by individuals at lower levels in the health system [125], this review harkens to the move to make health systems research a “people-centred science”, as “it is people who ultimately determine the character of a health system” in their capacities as users, providers, managers, knowledge agents, and financers [126]. It also recognises the importance of complexity in health systems research [17, 18]. This is consistent with recent calls in the health systems research literature to recognise that health systems are influenced by the settings in which they operate [126], and that health needs and outcomes are dynamic, evolving, and generated and shaped by social, economic, political, historical, and cultural forces [127].

Additionally, we did not apply language restrictions to our systematic review, which allowed us to search for, retrieve, and consider studies that were not published in English, thereby reducing language-related publication bias. We conducted searches of smaller, more regionally-focused databases (e.g. WPRIM), increasing our likelihood of capturing regional or smaller-scale research that may not have been accessible via or indexed in larger, standard academic databases.

This study is not without limitations. Heterogeneity of study designs, populations, analytical strategies, and effect measures meant it was not possible to conduct a meta-analysis. Included quantitative studies were of variable methodological quality, as evinced by the large proportion of studies rated as having an overall high risk of bias and and correspondingly small number of studies assessed to have low risk of bias across methodological domains. Although most included qualitative studies were found to have low risk of bias, it was not possible to draw causal inferences from them. As such, inferences about temporal and/or causal relationships between systems-level factors and diabetes outcomes could only be made for a limited number of factors, with careful consideration of context.

We are unable to exclude publication bias or reporting bias. For example, studies exploring effects of various factors on diabetes outcomes may have neglected to report results for health systems domains which did not achieve statistical significance. Furthermore, the review found that work from the United States was overrepresented, which provides helpful insights into gaps in the literature, but limits the applicability of our results to other contexts. Additionally, we found that limited evidence of studies assessing awareness as an outcome. Most awareness studies found in the initial search were focused on relationships between demographic and social factors (e.g. sex, age, educational level, income) on awareness, and not relationships between systems-level factors and awareness. Also, due to our focus on health systems factors and awareness, control, treatment, and adherence outcomes, it was not within this review’s scope to address broader questions around the sustainability (e.g. cost-effectiveness, cost-benefit) of interventions and programmes targeted at improving diabetes outcomes.

Policy implications

This review found an association between minimising out-of-pocket payments (e.g. drugs, self-monitoring equipment) and improved control and adherence in the North American context. This highlights the importance of reducing, or ideally eliminating, out of pocket costs of both prescription medication and monitoring supplies. It is unfortunate that few such studies were conducted elsewhere. There seems to be potential for moving to integrated care models where these do not yet exist, taking into account synergies between the dynamic and interacting components of integrated care models, including healthcare professionals aspects of care (e.g. education, empowerment, social support, case management), approaches to care delivery (e.g. structured care, multidisciplinary team care), and information technology (e.g. clinical decision support systems, shared electronic medical records).

Previous evidence has shown that task-sharing with non-physician healthcare workers improves management of non-communicable diseases [128]. Consistent with this, our review found evidence of an association between outcomes and sharing of certain care services and processes with pharmacists. However, it is not possible to rule out the influence of publication/reporting bias, and evidence remains dominated by studies from the United States.

Research implications

This study points to several possibilities for future research. Although we found evidence to support integrated, innovative care models, their implementation must be underpinned by a robust evidence base. As such, more studies looking at the impact of different types of care integration models on diabetes outcomes should be conducted. These studies may also generate valuable insights on which “mix” of care components and what types of integration are most conducive to achievement of positive outcomes.

There is also a need for more high-quality, longitudinal studies identifying the effect of health system arrangements on a variety of different outcomes. Studies of screening tended to be focused on uptake and adherence and not health and social outcomes. Additionally, we found diverse measures of medication adherence. There is a clear need for greater consistency here. Importantly, there were few studies focusing on upstream health system factors such as leadership and governance affecting good diabetes outcomes. For example, achieving integrated and innovated care delivery models is often dependent on effective management and lateral team management including different specialities, while expanding financial protection among larger sections of the population often results from political will; more research is needed on these associations.

Lastly, most studies are in high income countries. As the global prevalence of diabetes rises [129131] and the influence of health systems factors on chronic disease control and management is increasingly recognised [132134], efforts to generate evidence from low and middle-income countries (LMICs), especially considering their diverse population characteristics, diet and lifestyle shifts, sociocultural contexts, and health systems, should be augmented.

Supporting information

S1 Table. Study designs, settings, findings, and risk of bias of included studies.

https://doi.org/10.1371/journal.pone.0195086.s001

(DOCX)

S3 Text. Risk of bias assessment tool for observational studies.

https://doi.org/10.1371/journal.pone.0195086.s004

(DOCX)

S4 Text. Cochrane risk of bias assessment tool for randomised trials.

https://doi.org/10.1371/journal.pone.0195086.s005

(PDF)

S5 Text. Risk of bias assessment tool for qualitative studies.

https://doi.org/10.1371/journal.pone.0195086.s006

(DOCX)

S6 Text. Maimaris et al. 2013: The influence of health systems on hypertension awareness, treatment, and control: A systematic literature review.

https://doi.org/10.1371/journal.pone.0195086.s007

(PDF)

Acknowledgments

We thank Ms Mahin Sahih, Dr Miho Asano, and Ms Kiesha Prem for their assistance with foreign-language article translation.

References

  1. 1. International Diabetes Federation. IDF Diabetes Atlas—7th Edition. 2015 2015. Report No.
  2. 2. NCD Risk Factor Collaboration (NCD-RisC). Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4·4 million participants. The Lancet. 2016;387(10027):513–1530.
  3. 3. World Health Organization. Fact Sheet: Diabetes Geneva: World Health Organization; 2016 [updated June 2016; cited 2016 15 June]. Available from: https://web.archive.org/web/20130826174444/ http://www.who.int/mediacentre/factsheets/fs312/en/.
  4. 4. Chen L, Pei J-H, Kuang J, Chen H-M, Chen Z, Li Z, et al. Effect of lifestyle intervention in patients with type 2 diabetes: A meta-analysis. Metabolism. 2015;64:338–47. pmid:25467842
  5. 5. Baker M, Simpson K, Lloyd B, Bauman A, Fiatarone Singh M. Behavioral strategies in diabetes prevention programs: A systematic review of randomized controlled trials. Diabetes Research and Clinical Practice. 2011;91:1–12. pmid:20655610
  6. 6. van Dam R, Hu F. Coffee Consumption and Risk of Type 2 Diabetes: A Systematic Review. Journal of the American Medical Association. 2005;294(1):97–104. pmid:15998896
  7. 7. Cramer J. A Systematic Review of Adherence With Medications for Diabetes. Diabetes Care. 2004;27(5):1218–24. pmid:15111553
  8. 8. Krass I, Schieback P, Dhippayom T. Adherence to diabetes medication: a systematic review. Diabetic Medicine. 2015;32:725–37. pmid:25440507
  9. 9. Agardh E, Allebeck P, Hallqvist J, Moradi T, Sidorchuk A. Type 2 diabetes incidence and socio-economic position: a systematic review and meta-analysis. International Journal of Epidemiology 2011;40:804–18. pmid:21335614
  10. 10. Norris S, Nichols P, Caspersen C, Glasgow R, Engelgau M, Jack L, et al. The effectiveness of disease and case management for people with diabetes: a systematic review. American Journal of Preventive Medicine. 2002;22(4 Suppl):15–38.
  11. 11. Ali S, Giordano R, Lakhani S, Walker D. A review of randomized controlled trials of medical record powered clinical decision support system to improve quality of diabetes care. Int J Med Inform. 2016;87:91–100. pmid:26806716
  12. 12. Jeffery R, Iserman E, Haynes R, CDSS Systematic Review Team. Can computerized clinical decision support systems improve diabetes management? A systematic review and meta-analysis. Diabet Med. 2013;30(6):739–45. pmid:23199102
  13. 13. Balabanova D, McKee M, Koroleva N, Chikovani I, Goguadze K, et al. Navigating the health system: diabetes care in Georgia. Health Policy and Planning. 2009;24:46–54. pmid:19074492
  14. 14. Hopkinson B, Balabanova D, McKee M, Kutzin J. The human perspective on health care reform: coping with diabetes in Kyrgyzstan. International Journal of Health Planning and Management. 2004;19:43–61. pmid:15061289
  15. 15. World Health Organization. Health systems 2017 [cited 2017 12 October]. Available from: http://www.who.int/topics/health_systems/en/.
  16. 16. Gilson L, Lehmann U, Schneider H. Practicing governance towards equity in health systems: LMIC perspectives and experience. International Journal for Equity in Health. 2017;16(171).
  17. 17. Plsek P, Greenhalgh T. The challenge of complexity in health care. British Medical Journal. 2001;323(7313):625–8. pmid:11557716
  18. 18. Atun R. Health systems, systems thinking and innovation. Health Policy and Planning. 2012;27(4 Suppl 1):iv4–iv8.
  19. 19. Gilson L. Health policy and systems research: a methodology reader. Geneva: World Health Organization; 2012.
  20. 20. Maimaris W, Paty J, Perel P, Legido-Quigley H, Balabanova D, Nieuwlaat R, et al. The Influence of Health Systems on Hypertension Awareness, Treatment, and Control: A Systematic Literature Review. PLoS Medicine. 2013;10(7): e1001490. Epub 30 July 2013. pmid:23935461
  21. 21. Rees R, Harden A, Brunton G, Oliver S, Oakley A. Young people and physical activity: a systematic review of barriers and facilitators. London: EPPI-Centre, Social Science Research Unit, Institute of Education, University of London, 2001.
  22. 22. Harden A, Rees R, Shepherd J, Brunton G, S O, al. e. Young people and mental health: a systematic review of research on barriers and facilitators. London: EPPI-Centre, Social Science Research Unit, 2001.
  23. 23. Moher D, Liberati A, Tetzlaff J, Altman D. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. British Medical Journal. 2009;339:b2535. pmid:19622551
  24. 24. Babamoto KS, Sey KA, Camilleri AJ, Karlan VJ, Catalasan J, Morisky DE. Improving diabetes care and health measures among hispanics using community health workers: Results from a randomized controlled trial. Health Education and Behavior. 2009;36(1):113–26. pmid:19188371
  25. 25. Chao J, Yang L, Xu H, Yu Q, Jiang L, Zong M. The effect of integrated health management model on the health of older adults with diabetes in a randomized controlled trial. Archives of Gerontology and Geriatrics. 2015;60(1):82–8. pmid:25456892
  26. 26. Cohen LB, Taveira TH, Khatana SA, Dooley AG, Pirraglia PA, Wu WC. Pharmacist-led shared medical appointments for multiple cardiovascular risk reduction in patients with type 2 diabetes. Diabetes Educator. 2011;37(6):801–12. pmid:22021025.
  27. 27. Houweling ST, Kleefstra N, van Hateren KJ, Groenier KH, Meyboom-de Jong B, Bilo HJ. Can diabetes management be safely transferred to practice nurses in a primary care setting? A randomised controlled trial. Journal of Clinical Nursing. 2011;20(9–10):1264–72. pmid:21401764.
  28. 28. Rachmani R, Slavacheski I, Berla M, Frommer-Shapira R, Ravid M. Treatment of high-risk patients with diabetes: motivation and teaching intervention: a randomized, prospective 8-year follow-up study. Journal of the American Society of Nephrology. 2005;16 Suppl 1:S22–6. pmid:15938028.
  29. 29. Houweling ST, Kleefstra N, van Hateren KJ, Kooy A, Groenier KH, Ten Vergert E, et al. Diabetes specialist nurse as main care provider for patients with type 2 diabetes. Netherlands Journal of Medicine. 2009;67(7):279–84. pmid:19687522.
  30. 30. Jacobs M, Sherry PS, Taylor LM, Amato M, Tataronis GR, Cushing G. Pharmacist assisted medication program enhancing the regulation of diabetes (PAMPERED) study. Journal of the American Pharmacists Association. 2012;52(5):613–21. pmid:23023841
  31. 31. King AB, Wolfe GS. Evaluation of a diabetes specialist-guided primary care diabetes treatment program. Journal of the American Academy of Nurse Practitioners. 2009;21(1):24–30. pmid:19125892.
  32. 32. Krass I, Armour CL, Mitchell B, Brillant M, Dienaar R, Hughes J, et al. The Pharmacy Diabetes Care Program: assessment of a community pharmacy diabetes service model in Australia. Diabetic Medicine. 2007;24(6):677–83. pmid:17523968.
  33. 33. Odegard PS, Goo A, Hummel J, Williams KL, Gray SL. Caring for poorly controlled diabetes mellitus: A randomized pharmacist intervention. Annals of Pharmacotherapy. 2005;39(3):433–40. pmid:15701763
  34. 34. Ridgeway NA, Harvill DR, Harvill LM, Falin TM, Forester GM, Gose OD. Improved control of type 2 diabetes mellitus: a practical education/behavior modification program in a primary care clinic. Southern Medical Journal. 1999;92(7):667–72. pmid:10414474.
  35. 35. Taylor CT, Byrd DC, Krueger K. Improving primary care in rural Alabama with a pharmacy initiative. American Journal of Health-System Pharmacy. 2003;60(11):1123–9. pmid:12816022.
  36. 36. Davies MJ, Heller S, Skinner TC, Campbell MJ, Carey ME, Cradock S, et al. Effectiveness of the diabetes education and self management for ongoing and newly diagnosed (DESMOND) programme for people with newly diagnosed type 2 diabetes: Cluster randomised controlled trial. BMJ. 2008;336(7642):491–5. pmid:18276664
  37. 37. Edelman D, Dolor RJ, Coffman CJ, Pereira KC, Granger BB, Lindquist JH, et al. Nurse-led behavioral management of diabetes and hypertension in community practices: a randomized trial. Journal of General Internal Medicine. 2015;30(5):626–33. pmid:25567758.
  38. 38. Groeneveld Y, Petri H, Hermans J, Springer M. An assessment of structured care assistance in the management of patients with type 2 diabetes in general practice. Scandinavian Journal of Primary Health Care. 2001;19(1):25–30. pmid:11303543
  39. 39. Johnson JA, Majumdar SR, Bowker SL, Toth EL, Edwards A. Self-monitoring in Type 2 diabetes: A randomized trial of reimbursement policy. Diabetic Medicine. 2006;23(11):1247–51. pmid:17054603
  40. 40. Pladevall M, Divine G, Wells KE, Resnicow K, Williams LK. A randomized controlled trial to provide adherence information and motivational interviewing to improve diabetes and lipid control. Diabetes Educator. 2015;41(1):136–46. pmid:25486932.
  41. 41. McDermott RA, Schmidt B, Preece C, Owens V, Taylor S, Li M, et al. Community health workers improve diabetes care in remote Australian Indigenous communities: results of a pragmatic cluster randomized controlled trial. BMC Health Services Research. 2015;15(68).
  42. 42. Simpson SH, Majumdar SR, Tsuyuki RT, Lewanczuk RZ, Spooner R, Johnson JA. Effect of adding pharmacists to primary care teams on blood pressure control in patients with type 2 diabetes: a randomized controlled trial. Diabetes Care. 2011;34(1):20–6. pmid:20929988.
  43. 43. Yuan X, Wang F, Fish AF, Xue C, Chen T, Liu C, et al. Effect of case management on glycemic control and behavioral outcomes for chinese people with type 2 diabetes: A 2-year study. Patient Education and Counseling. 2016;99(8):1382–8. pmid:27005564
  44. 44. Heisler M, Vijan S, Makki F, Piette JD. Diabetes control with reciprocal peer support versus nurse care management: A randomized trial. Annals of Internal Medicine. 2010;153(8):507–15. pmid:20956707
  45. 45. Smith SM, Paul G, Kelly A, Whitford DL, O'Shea E, O'Dowd T. Peer support for patients with type 2 diabetes: cluster randomised controlled trial. BMJ. 2011;342:d715. pmid:21324992.
  46. 46. Heisler M, Hofer TP, Schmittdiel JA, Selby JV, Klamerus ML, Bosworth HB, et al. Improving blood pressure control through a clinical pharmacist outreach program in patients with diabetes mellitus in 2 high-performing health systems: the adherence and intensification of medications cluster randomized, controlled pragmatic trial. Circulation. 2012;125(23):2863–72. pmid:22570370.
  47. 47. Nielsen AB, de Fine Olivarius N, Gannik D, Hindsberger C, Hollnagel H. Structured personal diabetes care in primary health care affects only women's HbA1c.[Reprint in Ugeskr Laeger. 2007 Jun 18;169(25):2428–31; 17594836]. Diabetes Care. 2006;29(5):963–9. pmid:16644621.
  48. 48. Browning C, Chapman A, Yang H, Liu S, Zhang T, Enticott JC, et al. Management of type 2 diabetes in China: The Happy Life Club, a pragmatic cluster randomised controlled trial using health coaches. BMJ Open. 2016;6(3). pmid:26944692
  49. 49. Mshelia DS, Akinosun OM, Abbiyesuku FM. Effect of increased patient-physician contact time and health education in achieving diabetes mellitus management objectives in a resource-poor environment. Singapore Medical Journal. 2007;48(1):74–9. pmid:17245520.
  50. 50. Manders IG, Stoecklein K, Lubach CHC, Bijl-Oeldrich J, Nanayakkara PWB, Rauwerda JA, et al. Shift in responsibilities in diabetes care: The Nurse-Driven Diabetes In-Hospital Treatment protocol (N-DIABIT). Diabetic Medicine. 2016;33(6):761–7. pmid:26333117
  51. 51. Russell AW, Baxter KA, Askew DA, Tsai J, Ware RS, Jackson CL. Model of care for the management of complex Type 2 diabetes managed in the community by primary care physicians with specialist support: an open controlled trial. Diabetic Medicine. 2013;30(9):1112–21. pmid:23758279.
  52. 52. Bowker SL, Mitchell CG, Majumdar SR, Toth EL, Johnson JA. Lack of insurance coverage for testing supplies is associated with poorer glycemic control in patients with type 2 diabetes. CMAJ Canadian Medical Association Journal. 2004;171(1):39–43. pmid:15238494.
  53. 53. Burge MR, Lucero S, Rassam AG, Schade DS. What are the barriers to medical care for patients with newly diagnosed diabetes mellitus? Diabetes, Obesity and Metabolism. 2000;2(6):351–4. pmid:11225964
  54. 54. Parchman ML, Pugh JA, Wang CP, Romero RL. Glucose control, self-care behaviors, and the presence of the chronic care model in primary care clinics. Diabetes Care. 2007;30(11):2849–54. pmid:17682121
  55. 55. Piette JD, Heisler M, Krein S, Kerr EA. The role of patient-physician trust in moderating medication nonadherence due to cost pressures. Archives of Internal Medicine. 2005;165(15):1749–55. pmid:16087823
  56. 56. Piette JD, Heisler M, Wagner TH. Problems Paying Out-of-Pocket Medication Costs among Older Adults with Diabetes. Diabetes Care. 2004;27(2):384–91. pmid:14747218
  57. 57. Rhee MK, Cook CB, Dunbar VG, Panayioto RM, Berkowitz KJ, Boyd B, et al. Limited health care access impairs glycemic control in low income urban African Americans with type 2 diabetes. Journal of Health Care for the Poor & Underserved. 2005;16(4):734–46. pmid:16311495.
  58. 58. Sosa-Rubi SG, Galarraga O, Lopez-Ridaura R. Diabetes treatment and control: the effect of public health insurance for the poor in Mexico. Bulletin of the World Health Organization. 2009;87(7):512–9. pmid:19649365.
  59. 59. Kamien M, Ward AM, Mansfield F, Fatovich B, Mather C, Anstey K. Management of type 2 diabetes in Western Australian metropolitan general practice. Diabetes Research and Clinical Practice. 1994;26(3):197–208. pmid:7736900
  60. 60. Ngo-Metzger Q, Sorkin DH, Billimek J, Greenfield S, Kaplan SH. The effects of financial pressures on adherence and glucose control among racial/ethnically diverse patients with diabetes. Journal of General Internal Medicine. 2012;27(4):432–7. pmid:22005941
  61. 61. Tai TY, Chuang LM, Tsai ST, Huang BY. Treatment of type 2 diabetes mellitus in a primary care setting in Taiwan: Comparison with secondary/tertiary care. Journal of the Formosan Medical Association. 2006;105(2):105–17. pmid:16498504
  62. 62. Tan E, Yang W, Pang B, Dai M, Loh FE, Hogan P. Geographic variation in antidiabetic agent adherence and glycemic control among patients with type 2 diabetes. Journal of Managed Care and Specialty Pharmacy. 2015;21(12):1195–202. pmid:26679968
  63. 63. Chew B, Shariff-Ghazali S, Lee P, Cheong A, Mastura I, Haniff J, et al. Type 2 Diabetes Mellitus Patient Profiles, Diseases Control and Complications at Four Public Health Facilities- A Cross-sectional Study based on the Adult Diabetes Control and Management (ADCM) Registry 2009. Medical Journal of Malaysia. 2013;68(5):397–404. pmid:24632869
  64. 64. Drivsholm T, Olivarius NdF. General practitioners may diagnose type 2 diabetes mellitus at an early disease stage in patients they know well. Family Practice. 2006;23(3):192–7. pmid:16449242
  65. 65. Gibson TB, Song X, Alemayehu B, Wang SS, Waddell JL, Bouchard JR, et al. Cost sharing, adherence, and health outcomes in patients with diabetes. American Journal of Managed Care. 2010;16(8):589–600. pmid:20712392
  66. 66. Grogan M, Jenkins M, Sansing VV, MacGregor J, Brooks MM, Julien-Williams P, et al. Health insurance status and control of diabetes and coronary artery disease risk factors on enrollment into the bypass angioplasty revascularization investigation 2 diabetes (BARI 2D) trial. Diabetes Educator. 2010;36(5):774–83. pmid:20584997
  67. 67. Juul L, Maindal HT, Frydenberg M, Kristensen JK, Sandbaek A. Quality of type 2 diabetes management in general practice is associated with involvement of general practice nurses. Primary Care Diabetes. 2012;6(3):221–8. pmid:22554709
  68. 68. Littenberg B, Strauss K, MacLean CD, Troy AR. The use of insulin declines as patients live farther from their source of care: results of a survey of adults with type 2 diabetes. BMC Public Health. 2006;6:198. pmid:16872541.
  69. 69. Pilleron S, Pasquier E, Boyoze-Nolasco I, Villafuerte JJ, Olchini D, Fontbonne A. Participative decentralization of diabetes care in Davao City (Philippines) according to the Chronic Care Model: A program evaluation. Diabetes Research and Clinical Practice. 2014;104(1):189–95. pmid:24560175
  70. 70. Schmittdiel JA, Traylor A, Uratsu CS, Mangione CM, Ferrara A, Subramanian U. The association of patient-physician gender concordance with cardiovascular disease risk factor control and treatment in diabetes. Journal of women's health (2002). 2009;18(12):2065–70. pmid:20044871
  71. 71. Oyetayo OO, James C, Martinez A, Roberson K, Talbert RL. The Hispanic Diabetes Management Program: Impact of community pharmacists on clinical outcomes. Journal of the American Pharmacists Association. 2011;51(5):623–6. pmid:21896461
  72. 72. Schiel R, Voigt U, Ross IS, Braun A, Rillig A, Hunger-Dathe W, et al. Structured diabetes therapy and education improves the outcome of patients with insulin treated diabetes mellitus. The 10 year follow-up of a prospective, population-based survey on the quality of diabetes care (the JEVIN Trial). Experimental & Clinical Endocrinology & Diabetes. 2006;114(1):18–27. pmid:16450312.
  73. 73. de Sonnaville JJ, Bouma M, Colly LP, Deville W, Wijkel D, Heine RJ. Sustained good glycaemic control in NIDDM patients by implementation of structured care in general practice: 2-year follow-up study. Diabetologia. 1997;40(11):1334–40. pmid:9389427.
  74. 74. Elliott DJ, Robinson EJ, Anthony KB, Stillman PL. Patient-centered outcomes of a value-based insurance design program for patients with diabetes. Population Health Management. 2013;16(2):99–106. pmid:23405873
  75. 75. Gu Q, Zeng F, Patel BV, Tripoli LC. Part D coverage gap and adherence to diabetes medications. American Journal of Managed Care. 2010;16(12):911–8. pmid:21348561.
  76. 76. Spence MM, Makarem AF, Reyes SL, Rosa LL, Nguyen C, Oyekan EA, et al. Evaluation of an outpatient pharmacy clinical services program on adherence and clinical outcomes among patients with diabetes and/or coronary artery disease. Journal of managed care & specialty pharmacy. 2014;20(10):1036–45.
  77. 77. Kocarnik BM, Liu CF, Wong ES, Perkins M, Maciejewski ML, Yano EM, et al. Does the presence of a pharmacist in primary care clinics improve diabetes medication adherence? BMC Health Services Research. 2012;12:391. pmid:23148570.
  78. 78. Hsu J, Price M, Huang J, Brand R, Fung V, Hui R, et al. Unintended consequences of caps on medicare drug benefits. New England Journal of Medicine. 2006;354(22):2349–59. pmid:16738271
  79. 79. Hull S, Chowdhury TA, Mathur R, Robson J. Improving outcomes for patients with type 2 diabetes using general practice networks: A quality improvement project in east London. BMJ Quality and Safety. 2014;23(2):171–6. pmid:24003237
  80. 80. Hunt J, Rozenfeld Y, Shenolikar R. Effect of patient medication cost share on adherence and glycemic control. Managed care (Langhorne, Pa). 2009;18(7):47–53.
  81. 81. Panarotto D, Gravina LB, Teles AR, Oliveira MSd, Trãsel HdAV. Controle glicêmico de pacientes diabéticos tipo 2 nos serviços público e privado de Saúde. Arq Bras Endocrinol Metabol. 2009;53(6):733–40. pmid:19893916
  82. 82. Pawaskar M, Burch S, Seiber E, Nahata M, Iaconi A, Balkrishnan R. Medicaid payment mechanisms: Impact on medication adherence and health care service utilization in type 2 diabetes enrollees. Population Health Management. 2010;13(4):209–18. pmid:20455787
  83. 83. Salinas-Martinez AM, Garza-Sagastegui MG, Cobos-Cruz R, Nunez-Rocha GM, Garza-Elizondo ME, Peralta-Chavez DF. [Effects of incorporating group visits on the metabolic control of type 2 diabetic patients]. Revista Medica de Chile. 2009;137(10):1323–32. pmid:20011939.
  84. 84. Tranche S, Galgo A, Mundet X, Sanchez-Zamorano MA, Karen APSI. Cardiovascular risk factors in type 2 diabetic patients: multifactorial intervention in primary care. Kidney International—Supplement. 2005;(93):S55–62. pmid:15613070.
  85. 85. Ziemer DC, Miller CD, Rhee MK, Doyle JP, Watkins C Jr., Cook CB, et al. Clinical inertia contributes to poor diabetes control in a primary care setting. Diabetes Educator. 2005;31(4):564–71. pmid:16100332.
  86. 86. Chalermsri C, Paisansudhi S, Kantachuvesiri P, Pramyothin P, Washirasaksiri C, Srivanichakorn W, et al. The effectiveness of holistic diabetic management between Siriraj Continuity of Care Clinic and medical out-patient department. Journal of the Medical Association of Thailand. 2014;97(3 SUPPL. 3):S197–S205.
  87. 87. Bunting BA, Lee G, Knowles G, Lee C, Allen P. The hickory project: Controlling healthcare costs and improving outcomes for diabetes using the Asheville project model. American Health and Drug Benefits. 2011;4(6):343–50. pmid:25126361
  88. 88. Bello SI, Ganiyu KA, Dakop YO, Erah PO. Pharmacist's intervention in the control of blood sugar levels in randomised diabetes patients at a primary health care setting in Benin City. Nigerian Quarterly Journal of Hospital Medicine. 2012;22(4):245–8. pmid:24568058.
  89. 89. Berdine HJ, Skomo ML. Development and integration of pharmacist clinical services into the patient-centered medical home. Journal of the American Pharmacists Association: JAPhA. 2012;52(5):661–7. pmid:23023848.
  90. 90. Kiblinger L, Braza NL. The Impact of Diabetes Education on Improving Patient Outcomes. Insulin. 2007;2(1):24–30.
  91. 91. Monte SV, Slazak EM, Albanese NP, Adelman M, Rao G, Paladino JA. Clinical and economic impact of a diabetes clinical pharmacy service program in a university and primary care-based collaboration model. Journal of the American Pharmacists Association: JAPhA. 2009;49(2):200–8. pmid:19289346
  92. 92. Patel AA, Kuti EL, Dale KM, Shah SA, White CM, Coleman CI. Effect of a medication assistance program on clinical outcomes in patients with diabetes. Formulary. 2006;41(10):518–22.
  93. 93. Reed RL, Revel AO, Carter A, Saadi HF, Dunn EV. A clinical trial of chronic care diabetic clinics in general practice in the United Arab Emirates: A preliminary analysis. Archives of Physiology and Biochemistry. 2001;109(3 SPEC ISS.):272–80. pmid:11880932
  94. 94. Coast-Senior EA, Kroner BA, Kelley CL, Trilli LE. Management of patients with type 2 diabetes by pharmacists in primary care clinics. Annals of Pharmacotherapy. 1998;32(6):636–41. pmid:9640480
  95. 95. Coleman K, Reiter KL, Fulwiler D. The impact of pay-for-performance on diabetes care in a large network of community health centers.[Erratum appears in J Health Care Poor Underserved. 2008 May;19(2):657]. Journal of Health Care for the Poor & Underserved. 2007;18(4):966–83. pmid:17982218.
  96. 96. Ko I, Lee T, Kim G, Kang S, Kim MJ. Effects of visiting nurses' individually tailored education for low-income adult diabetic patients in Korea. Public Health Nursing. 2011;28(5):429–37. pmid:22092426
  97. 97. Musacchio N, Giancaterini A, Lovagnini Scher A, Errichelli C, Pessina L, Salis G, et al. Impact of a chronic care model based on patient empowerment on the management of type 2 diabetes: Effects of the SINERGIA program. Giornale Italiano di Diabetologia e Metabolismo. 2010;30(2):49–55.
  98. 98. Vella J, Azzopardi LM. Pharmacist-led diabetic patient monitoring. Journal of EuroMed Pharmacy. 2013;(3):21–4.
  99. 99. Kengne AP, Fezeu L, Sobngwi E, Awah PK, Aspray TJ, Unwin NC, et al. Type 2 diabetes management in nurse-led primary healthcare settings in urban and rural Cameroon. Primary Care Diabetes. 2009;3(3):181–8. pmid:19748331
  100. 100. Ko JJ, Lu J, Rascati K, Stock EM, Juan J, Suh K, et al. Analysis of Glycemic Control of a Pharmacist-Led Medication Management Program in Patients with Type 2 Diabetes. Journal of Managed Care & Specialty Pharmacy. 2016;22(1):32–7. pmid:27015049.
  101. 101. Spigt M, Rikkers A, Doornbos M, Wouters E, Spitz I, Van Amelsvoort L, et al. The effect of screening on the prevalence of diagnosed type 2 diabetes in primary care. Scandinavian Journal of Primary Health Care. 2009;27(4):232–7. pmid:19929181
  102. 102. Quinones ME, Pio MY, Chow DH, Moss E, Hulstein JL, Boatright SM, et al. Impact of clinical pharmacy services on outcomes and costs for indigent patients with diabetes. American Journal of Managed Care. 2016;22(4):e147–e52. pmid:27143351
  103. 103. Hueston WJ. Does having a personal physician improve quality of care in diabetes? Journal of the American Board of Family Medicine: JABFM. 2010;23(1):82–7. pmid:20051546.
  104. 104. Maschak-Carey BJ, Schutta MH, Schwartz SS, Joshi MS, Shulkin DJ, Bernard DB. Successful implementation of a comprehensive diabetes disease management program in an integrated academic health system. Disease Management. 1999;2(1–2):43–50.
  105. 105. Pinsky B, Harnett J, Paulose-Ram R, Mardekian J, Samant N, Nair KV. Impact of treatment by NCQA-certified physicians on Diabetes-Related outcomes. American Health and Drug Benefits. 2011;4(7).
  106. 106. Richardson GC, Derouin AL, Vorderstrasse AA, Hipkens J, Thompson JA. Nurse practitioner management of type 2 diabetes. Permanente Journal. 2014;18(2):e134–40. pmid:24867560.
  107. 107. Soumerai SB, Mah C, Zhang F, Adams A, Barton M, Fajtova V, et al. Effects of Health Maintenance Organization Coverage of Self-monitoring Devices on Diabetes Self-care and Glycemic Control. Archives of Internal Medicine. 2004;164(6):645–52. pmid:15037493
  108. 108. Lewis CP, Newell JN. Patients' perspectives of care for type 2 diabetes in Bangladesh -a qualitative study. BMC public health. 2014;14:737. pmid:25047797
  109. 109. Alberti H, Boudriga N, Nabli M. Primary care management of diabetes in a low/middle income country: A multi-method, qualitative study of barriers and facilitators to care. BMC Family Practice. 2007;8. pmid:17996084
  110. 110. Bhojani U, Mishra A, Amruthavalli S, Devadasan N, Kolsteren P, De Henauw S, et al. Constraints faced by urban poor in managing diabetes care: patients' perspectives from South India. Global health action. 2013;6:22258. pmid:24093885
  111. 111. Gazmararian JA, Ziemer DC, Barnes C. Perception of barriers to self-care management among diabetic patients. Diabetes Educator. 2009;35(5):778–88. pmid:19556552
  112. 112. Jeragh-Alhaddad FB, Waheedi M, Barber ND, Brock TP. Barriers to medication taking among Kuwaiti patients with type 2 diabetes: A qualitative study. Patient Preference and Adherence. 2015;9:1491–503. pmid:26604702
  113. 113. Mendenhall E, Norris SA. Diabetes care among urban women in Soweto, South Africa: a qualitative study. BMC Public Health. 2015;15(1300).
  114. 114. Collinsworth AW, Vulimiri M, Schmidt KL, Snead CA. Effectiveness of a Community Health Worker-led Diabetes Self-Management Education Program and Implications for CHW Involvement in Care Coordination Strategies. Diabetes Educator. 2013;39(6):792–9. pmid:24052203
  115. 115. Katz I, Schneider H, Shezi Z, Mdleleni G, Gerntholtz T, Butler O, et al. Managing type 2 diabetes in Soweto-The South African Chronic Disease Outreach Program experience. Primary Care Diabetes. 2009;3(3):157–64. pmid:19640820
  116. 116. Loskutova NY, Tsai AG, Fisher EB, LaCruz DM, Cherrington AL, Harrington TM, et al. Patient Navigators Connecting Patients to Community Resources to Improve Diabetes Outcomes. Journal of the American Board of Family Medicine: JABFM. 2016;29(1):78–89. pmid:26769880.
  117. 117. Ouwens M, Wollersheim H, Hermens R, Hulscher M, Grol R. Integrated care programmes for chronically ill patients: a review of systematic reviews. International Journal for Quality in Health Care. 2005;17(2):141–6. pmid:15665066
  118. 118. Renders C, Wagner E, Valk G, Eijk J, Griffin S, Assendelft W. Interventions to Improve the Management of Diabetes in Primary Care, Outpatient, and Community Settings: a systematic review. Diabetes Care. 2001;24:1821–33. pmid:11574449
  119. 119. Kodner D, Spreeuwenberg C. Integrated care: meaning, logic, applications, and implications–a discussion paper. International Journal of Integrated Care. 2002;2:e12. pmid:16896389
  120. 120. Kodner D. All together now: a conceptual exploration of integrated care. Healthcare Quarterly. 2009;13(Spec):6–15. pmid:20057243
  121. 121. Wulsin L, Sollner W, Pincus H. Models of integrated care. Medical Clinics of North America. 2006;90:647–77. pmid:16843767
  122. 122. World Health Organization Regional Office for Europe. Integrated care models: an overview—working document. Geneva, Switzerland: World Health Organization Regional Office for Europe, 2016.
  123. 123. Machado M, Bajcar J, Guzzo G, Einarson T. Sensitivity of patient outcomes to pharmacist interventions. Part I: systematic review and meta-analysis in diabetes management. Annals of Pharmacotherapy. 2007;41(10):1569–82. pmid:17712043
  124. 124. Santschi V, Chiolero A, Burnand B, Colosimo A, Paradis G. Impact of pharmacist care in the management of cardiovascular disease risk factors: a systematic review and meta-analysis of randomized trials. Archives of Internal Medicine. 171(16):1441–53. pmid:21911628
  125. 125. Pyone T, Smith H, van den Broek N. Frameworks to assess health systems governance: a systematic review. Health Policy and Planning. 2017;32(5):710–22. pmid:28334991
  126. 126. Sheikh K, George A, Gilson L. People-centred science: strengthening the practice of health policy and systems research. BMC Health Research Policy and Systems. 2014;12(19):1–18.
  127. 127. Balabanova D, Mills A, Conteh L, Akkazieva B, Banteyerga H, Dash U, et al. Good health at low cost 25 years on: lessons for the future of health systems strengthening. The Lancet. 2013;381(9883):2118–33.
  128. 128. Joshi R, Alim M, Kengne A, Jan S, Maulik P, Peiris D, et al. Task Shifting for Non-Communicable Disease Management in Low and Middle Income Countries–A Systematic Review. PloS One. 2014;9(8):e103754. pmid:25121789
  129. 129. Hu F. Globalization of diabetes: the role of diet, lifestyle, and genes. Diabetes Care. 2011;34(6):1249–57. pmid:21617109
  130. 130. Yoon K, Lee J, Kim J, Cho J, Choi Y, Ko S, et al. Epidemic obesity and type 2 diabetes in Asia. The Lancet. 2006;368:1681–88.
  131. 131. Abuyassin B, Laher I. Diabetes epidemic sweeping the Arab world. World Journal of Diabetes. 2016;7(8):165–74 pmid:27114755
  132. 132. World Health Organization. Global Action Plan for the Prevention and Control of NCDs 2013–2020. Geneva, Switzerland: World Health Organization, 2013 2013. Report No.
  133. 133. Atun R, Jaffar S, Nishtar S, Knaul F, Barreto M, Nyirenda M, et al. Improving responsiveness of health systems to non-communicable diseases. The Lancet. 2013;381(9867):609–97.
  134. 134. Samb B, Desai N, Nishtar S, Mendis S, Bekedam H, Wright A, et al. Prevention and management of chronic disease: a litmus test for health-systems strengthening in low-income and middle-income countries. The Lancet. 2010;376(9754):1785–97.