Social deprivation and foot disease outcomes in people with Diabetes Mellitus and Chronic Kidney Disease: A Systematic Reviews and Meta-Analysis Protocol

Primary question

Is there an association between higher levels of social deprivation and foot disease outcomes in individuals with diabetes and chronic kidney disease?

Objectives

Diabetes-related foot disease

The consensus definition from the International Working Group on the Diabetic Foot (IWGDF) refers to conditions affecting the foot of an individual with a past or present diagnosis of Diabetes Mellitus and includes one or more of the following²⁰:

Diabetes-related Foot Ulcer (DFUs)

Recognised by the IWGDF, DFUs are severe complications of DM, defined as break(s) of the skin at the foot that can range from superficial involvement of the epidermis to deeper penetration into the dermis, in individuals with a past or present diagnosis of Diabetes Mellitus. It is usually associated with loss of protective sensation (LOPS), foot deformity, and/or peripheral artery disease in the lower extremities.

Secondary outcome measures investigating the severity of DFU: Certain classification systems will be used. In line with the IWGDF Guidelines, the classification system used for measuring the severity of DFU are⁶;

The IWGDF strongly recommends six out of 28 foot disease classification systems—DIAFORA, IDSA/IWGDF, SINBAD, University of Texas Classification System (UTWCS), Wagner, and WIFi— because of their accuracy, moderate desirable effects, and minimal undesirable effects, making them suitable for clinical practice. Despite the low certainty of the evidence, these systems were favoured for their overall balance of effects. Although accepted for this study, DIAFORA, UTWCS, and Wagner were not selected for specific clinical use or audits because of limitations, such as lower reliability, applicability, or comprehensiveness in certain clinical scenarios⁶.

Amputations

Mortality

Mortality is directly or indirectly associated with the development of diabetes-related foot complications²¹.

Diabetes-related foot disease hospitalisation

Admission of an inpatient to a hospital primarily or secondarily due to diabetes-related foot disease²¹.

Chronic kidney disease

Renal or kidney damage can arise from both microvascular damage (diabetic kidney disease) and arterial damage (renovascular damage)²². In this study, we will characterise CKD as abnormalities of kidney structure or function with a sustained decrease in estimated glomerular filtration rate (eGFR) ≥59 or urine albumin-creatinine ratio (uACR)) >3 subsequent to hyperglycemia on two occasions for greater than three months²³.

It is categorised into five stages determined by GFR (G1-G5), and albuminuria category (A1-A3). The first stage is indicated by albuminuria or other abnormalities in kidney function, whereas the remaining four stages are marked by reductions in the GFR to below 90, 60, 30, and 15 ml/minute/ 1.73m² (Table 1)²³.

Social deprivation

Social deprivation is conceptualised and operationalised in different ways internationally. Townsend regarded deprivation as a barrier to resources such as ‘diet, clothing, housing, household facilities and fuel and environmental, education, working and social conditions, activities, and facilities, which are deemed necessities and heavily hinder an individual’s living conditions²⁴.

The foundation for multiple deprivation was adapted from Townsend^24,25. It considers seven main domains: income, employment, education, health, crime, access to housing and services, and living environment, where low income is a proxy for material deprivation^24,26.

Therefore, we defined social deprivation as the reduction or absence of social and economic resources, opportunities, and access to essential services. It is not limited to a single factor but encompasses multiple dimensions, such as ethnicity, unemployment, poor education and skills, low income, poor housing, living environment, social exclusion, crime, family breakdown, and barriers to healthcare/service that hinder the physical and mental health of individuals.

While there are different strata of social deprivation, some individuals may experience one form of deprivation, whereas others may face multiple or particularly severe forms of social deprivation, making them more likely to be exposed to the broader impacts of social deprivation. Hence, because of the complexity and multifaceted nature of social deprivation, two or more dimensions must be considered when measuring it to capture some of its numerous interconnected characteristics and their impact on inequalities, disadvantages, and discrimination²⁷.

Population

Adults aged 18 years and older Individuals diagnosed with DM (type 1, type 2, or other) and CKD (regardless of the CKD stage). There were no restrictions based on sex, race, or ethnicity. To measure the prevalence and incidence rates of diabetes-related foot disease (DFD), including DFU, nontraumatic amputation, and mortality, we will include individuals with and without DFD. Studies will be excluded if they do not focus on patients with CKD, DM, and DFD or if they focus on other comorbidities without separate analysis for DM and CKD. Additionally, studies were excluded if CKD was not a consequence of DM.

Exposure/Intervention

Studies reporting on the influence of social deprivation factors, including but not limited to 1) socioeconomic status (SES), 2) poverty, 3) Low Income, and 4) deprivation index or similar measures. Using validated measures or indices, such as the index of multiple deprivation or socioeconomic status scales, such as Plobal HP, Townsend, or Carstairs. Studies that do not assess the complexity of social deprivation and/or use one variable (household income) as a proxy for measuring social deprivation will be excluded.

Comparison/Control

Studies investigating the influence of social deprivation on foot disease outcomes have associated exposure to social disadvantages among individuals with DM and CKD. First, this will include observational studies, given the nature of the primary aim. Studies without a clear comparison of groups or gradient of social deprivation were excluded. In addition, studies that do not provide a comparative analysis or report only a single level of deprivation without comparison.

Outcome

Endpoints that are critical for decision-making are of primary interest. The primary outcomes have been prioritised based on their direct impact on patient health and the severity of complications associated with diabetes-related foot disease in individuals with CKD. We will prioritise the analysis and grading of the primary outcomes: 1) incidence and prevalence of foot ulcers, 2) frequency of non-traumatic lower extremity amputations (both minor and major), and 3) mortality.

The secondary outcomes provide additional insights into the burden on the healthcare system and the effectiveness of management strategies. Secondary outcomes of interest included 1) Severity of DFU, 2) hospital admission (duration of hospital stay), 3) frequency of lower limb health interventions and care, and 4) healing time for diabetes-related foot ulcers (DFUs) with a focus on infection. These secondary outcomes provide additional insights into the broader implications of diabetes-related foot complications on healthcare resources and patients’ quality of life. Studies that did not report foot disease outcomes or only reported general health outcomes unrelated to foot disease were excluded. Studies with outcomes that are not clearly defined, measured, or that use validated measurement tools will be excluded.

Across all healthcare settings, including hospitals, outpatient clinics, and community settings, the investigation primarily included observational studies (e.g., cohort studies, case-control studies, and cross-sectional studies). Additionally, interventional studies (e.g., randomised controlled trials, quasi-experimental studies), systematic reviews, and meta-analyses relevant to the primary and secondary outcomes will be considered, as well as grey literature. There will be no restriction on geographical location to capture potential differences in the social determinants of the impact of health and healthcare systems on foot disease outcomes. The review will include studies published between the retrospective starting date and search. Further eligibility criteria require that studies must meet a minimum quality threshold based on a standardised assessment tool, including the MOOSE Guidelines for meta-analyses in observational studies and the JBI Manual for Evidence Synthesis^30,31.

Information sources

We will use sensitive topic-based strategies tailored for each database, searching each one from its starting date until the commencement of our search.

Medical databases

MEDLINE (OVID interface), University of Galway Library, PubMed, Embase, Cochrane Library, CINAHL, Scopus Social.

Science database

PsycINFO, SocINDEX

Grey literature sources

OpenGrey, ProQuest Dissertations and Theses Global, Google Scholar, ClinicalTrials.gov, WHO Global Health Library (LILACS, AIM)

Search strategy

The draft of the MEDLINE (Ovid Interface) search was adapted from the general approach described by Hausner et al. and Chapter 4 of the Cochrane Guidelines^32,33. Search terms were added to the social deprivation-focused search strategy to improve sensitivity, whereas others were removed to increase specificity. The draft search strategy was peer-reviewed, updated, and validated towards the end of the review to ensure that the MEDLINE strategy effectively yielded a high proportion of eligible studies.

Study design

Data management

To manage the records and data throughout the systematic review, we implemented a structured approach using several tools and processes. Initially, comprehensive literature searches will be conducted across multiple databases, such as Endnote and Covidence databases. Endnotes will be used to organise and store all extracted data and related documentation. Covidence will be used to manage screening and data extraction processes, allowing for blinded and independent review by reviewers and tracking decisions and justifications for inclusion or exclusion of studies.

Risk of bias

Our approach involves assessing the risk of bias at both the study and outcome levels using established tools: the MOOSE Guidelines for meta-analysis in Observational studies and the JBI Manual for Evidence Synthesis^30,31. If details are insufficient, bias will be rated as ‘unclear,’ and the study investigators will be contacted for clarification. These judgments will be made independently by the two review authors, with disagreements resolved through discussion with a third reviewer. Visual representations of potential biases within and across studies were generated using RevMan 5.1 (Review Manager 5.1). Each domain in the risk of bias assessment was considered independently without an overall score.

To evaluate publication bias across the included studies, funnel plots were used when at least 10 studies were available for a particular outcome. An asymmetrical funnel plot suggests the presence of publication bias. Additionally, we will employ statistical tests, such as Egger’s regression test, to formally assess the presence of publication bias. To further mitigate publication bias, we will search for gray literature, including conference abstracts, theses, and reports, to include studies that might not be published in peer-reviewed journals. However, we critically appraised the quality and methodology of these studies, acknowledging the inherent limitations associated with grey literature, such as variability in study design and quality. We will also assess for any reporting bias by comparing findings from gray literature with those from peer-reviewed sources.

We will compare the outcomes reported in published studies with the outcomes listed in their protocols or trial registries, if available. Discrepancies were noted as potential indicators of selective reporting biases. We will also check for the completeness of outcome reporting by assessing whether all prespecified outcomes are reported in the study findings and addressing potential biases specific to observational research methodologies.

We will use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework to assess the quality of evidence across the included studies^34,35. Each outcome included in the PICO will be evaluated for certainty using GRADE criteria. Evidence begins with a high-quality rating for randomized trials and a low-quality rating for observational studies, which may be downgraded based on factors such as the risk of bias, inconsistency, indirectness, imprecision, and publication bias. Alternatively, evidence may be upgraded if factors such as a large effect size, a clear dose-response relationship, or indications that plausible confounding would reduce the strength of an effect suggest that the true effect may be greater than initially estimated.

We conducted sensitivity analyses to assess the robustness of our findings. This will include re-running analyses by excluding studies with a high risk of bias, or those identified as outliers. Additionally, we used the trim-and-fill method to adjust for publication bias.

Data synthesis

Residual heterogeneity will be evaluated using the I² statistic and the corresponding chi-square tests. Outcomes will be combined and calculated using RevMan 5.1, and calculations will adhere to the Cochrane Handbook for Systematic Review of Interventions Guidelines to ensure standardised and rigorous methods³³. If there was no significant heterogeneity among the studies, the Mantel-Haenszel method was used within a fixed effect model. If statistical heterogeneity is observed, indicated by an I² statistic of ≥50% or a P-value of <0.1, a random-effects model will be used. In cases where substantial heterogeneity is present (I² >75%), a meta-analysis will not be conducted because of the high variability among the study results. Instead, a qualitative narrative synthesis is used to provide a comprehensive summary of the findings.

I² values will be interpreted as follows

If high levels of heterogeneity are observed (I² statistic of ≥50% or a P-value of <0.1), we will analyse the study design and characteristics of the included studies to identify potential sources of variability. Sensitivity and subgroup analyses were conducted to explore the impact of varying definitions and measures on the results.

In cases where quantitative synthesis is inappropriate due to substantial heterogeneity in the study population, intervention, comparison, outcomes, or study methodology, we will use qualitative narrative synthesis to summarise the findings. Narrative synthesis is structured to convey a comprehensive understanding of the data collected, even when statistical aggregation is not feasible. In line with the guidelines of the Centre for Reviews and Dissemination³⁶, we will explore the relationship and findings both within and across the included studies.

The results are presented according to the key questions addressed in this review. For each key question, the primary outcomes are discussed before the secondary outcomes. Information will be prioritised to reflect the overall patient groups first, followed by subgroups defined by sociodemographic factors (e.g., age, sex, ethnicity, socioeconomic status) and clinical characteristics (e.g., type 1 DM or Type 2 DM and severity of CKD). Primary health outcomes (e.g., rates of lower-extremity amputation and mortality) will be prioritised over intermediate outcomes. Patient-related outcomes directly affecting patient health and well-being will be discussed before the utilisation outcomes (e.g., previous intervention and duration of hospital admissions).

Studies with a high risk of bias will be identified and discussed separately from those with a low or moderate risk. We will highlight the studies retained for certain key questions or outcomes based on their risk of bias and methodological quality.

Given the nature of the primary aim, results from observational studies will be presented first, supplemented by data from randomised controlled trials where available. Information will be stratified based on key methodological aspects, such as the blinding of investigators, patients, and outcome assessors. Reliable and valid instruments used to measure outcomes (e.g., validated tools for assessing foot ulcer severity) are emphasised.

Tables and diagrams will be used extensively to summarise the characteristics of the studies, focusing on those with a low or moderate risk of bias, and principal comparisons and key outcomes will be tabulated to provide a clear overview.

We will conduct subgroup analyses based on key covariates, including patient characteristics (such as age, sex, and ethnicity), differentiation between type 1 DM and Type 2 DM, stages of CKD (with consideration of grouping stages into broader categories if necessary), and DFU severity (mild, moderate, severe). Social deprivation levels will also be assessed using validated indices to distinguish between high and low socioeconomic status. Additionally, interventions will be categorized into surgical, vascular, podiatric, wound care, and offloading. Each subgroup will be clearly defined, and a random-effects model will be used to accommodate the variability within and between subgroups. Subgroup analysis will be conducted independently to explore variability based on predefined criteria, such as patient characteristics (age, sex, socioeconomic status), type of intervention, type of diabetes (type 1 DM vs. type 2 DM), and stages of chronic kidney disease (early stage CKD vs. late-stage CKD).

Dichotomous outcomes

For Dichotomous outcomes (presence or absence of DFU/amputation/infection, whether a DFU has healed completely, whether a patient has been hospitalised due to diabetes-related foot complications), we will calculate the risk ratios (RRs) with 95% confidence intervals (CIs) for each study. We will combine the results from individual studies using a random-effects model to account for the variability among studies. If heterogeneity was low (I2 <50%), a fixed-effect model was used.

Continuous outcomes

For Continuous outcomes (severity of foot ulcers measured by a specific classification system or the duration of ulcer healing), we will calculate the mean difference (MDs) or standardised mean difference (SMDs) with 95% confidence intervals, depending on whether the outcomes are measured using the same or different scales across the studies. The results were pooled using a random-effects model and a fixed-effects model if heterogeneity was low.