Abstract
This study evaluated the effectiveness of a multi‐disciplinary diabetic limb salvage programme in improving clinical outcomes and optimising healthcare utilisation in 406 patients aged ≥80 years with diabetic foot ulcers (DFUs), compared to 2392 younger patients enrolled from June 2020 to June 2021 and against 1716 historical controls using one‐to‐one propensity score matching. Results showed that elderly programme patients had lower odds of amputation‐free survival (odds ratio: 0.64, 95% CI: 0.47, 0.88) and shorter cumulative length of stay (LOS) compared to younger programme patients (incidence rate ratio: 0.45, 95% CI: 0.29, 0.69). Compared to the matched controls, participating in the programme was associated with 5% higher probability of minor lower extremity amputation, reduced inpatient admissions and emergency visits, shorter LOS but increased specialist and primary care visits (all p‐values <0.05). The findings suggest that the programme yielded favourable impacts on the clinical outcomes of patients aged≥80 years with DFUs. Further research is needed to develop specific interventions tailoring to the needs of the elderly population and to determine their effectiveness on patient outcomes while accounting for potential confounding factors.
Keywords: amputation, diabetic foot ulcer, health impact assessment, interdisciplinary communication, lower extremity
Abbreviations
- ATEs
Average treatment effects
- CKD
Chronic kidney disease
- DFUs
Diabetes foot ulcers
- DLS
Diabetic limb salvage
- DM
Diabetes Mellitus
- DS
Day surgery
- IHD
Ischemic heart disease
- LEA
Lower extremity amputation
- LEAPP
Lower Extremity Amputation Prevention Programme
- LOS
Length of stay
- SOC
Specialist outpatient clinic
- STEP
Screening and Surveillance, Treatment, Escalation Programme for ulcer prevention
- MDT
Multi‐disciplinary team
- PSM
Propensity score matching
1. INTRODUCTION
Diabetes Mellitus (DM) represents a substantial public health concern both globally and in South‐East Asia. Worldwide, the age‐standardised prevalence of diabetes in the adult population is projected to increase from 9.3% in 2019 to 10.2% by 2030, and further to 10.9% by 2045. In South‐East Asia, this prevalence is projected to increase from 11.3% in 2019 to 12.2% by 2030, and further to 12.6% by 2045. 1 In Singapore, the crude prevalence of type 2 diabetes in adults aged 18–69 years is estimated to rise from 8.6% in 2017 2 to 15.0% in 2050, 3 emphasising the growing significance of this issue. Diabetic foot ulcer (DFU) is one of the most common and serious complications of diabetes and is associated with an elevated risk of amputation and mortality. 4 Diabetes‐related lower extremity amputation (LEA), typically considered as the last‐resort treatment for patients with an uncontrollable severe foot infection in diabetic foot ulceration, is associated with significant functional decline and impaired quality of life, causing substantial economic burden to the healthcare system. Therefore, reducing LEAs is a primary marker of quality of care for the diabetic foot. 5 In contrast to a decreasing trend in other countries, 6 the rate of diabetes‐related LEA in Singapore was relatively stable between 2008 and 2017 and has remained one of the countries with the highest rates of LEA in the World. 7
Older age is one of the known risk factors for adverse outcomes in peripheral artery disease, a major cause of DFUs in patients with diabetes. 8 Singapore has a rapidly ageing population, with citizens aged 65 years and above accounting for 18.4%, 9 and the life expectancy at birth was 83.0 years in 2022. 10 With the ongoing ageing of the population and the anticipated rise in diabetes prevalence among older adults, the total number of elderly individuals with DFUs necessitating diabetes‐related LEA may increase. Some clinicians may advocate primary amputation or palliative wound management, as opposed to diabetic limb salvage (DLS) in elderly DFU patients with severe diabetes complications. 11 However, older adults, especially those aged ≥80 years, are particularly susceptible to a range of pathological factors that significantly influence their fitness to undergo and recover from LEA, resulting in an increased post‐amputation mortality rate. 12 Hence, it is imperative to design tailored and effective treatment approaches to address the unique challenges and risks associated with amputation in the elderly population, which ultimately contribute to better outcomes and quality of life.
In recent years, with improved understanding of the aetiology and underlying pathophysiology of DFUs, the general management approaches have been extensively reviewed. There has been a growing focus on the development and implementation of multi‐disciplinary care coordination programmes aimed at DFU management worldwide. 13 , 14 , 15 DLS, a highly complex procedure involving various techniques to remove any dead tissue and replace devitalised infected bone so that the lower limb can be preserved as much as possible with reduced risk of future infections and can function properly for ambulation, should be considered as a first‐line approach in treating patients with DFUs. 16 In Singapore, a LEA prevention programme implemented in tertiary care clinics has demonstrated encouraging results in reducing the incidence of both minor and major LEA, as well as improving cardiovascular health by improved glycaemic and lipid control. 17 , 18 Nevertheless, the effect of patients' age on LEA outcomes and healthcare utilisation as well as the effects of a multi‐disciplinary DLS programme in individuals aged ≥80 years have not been investigated. Thus, this study aimed to examine the association between age group (individuals aged ≥80 years and those aged <80 years old) and the outcomes and evaluate the effectiveness of the DLS programme in reducing LEA and mortality rates and optimising healthcare utilisation among individuals aged ≥80 years as compared to a historical control group. The selection of 80 years old as the cut‐off age is based on clinical expert opinion as well as literature showing that older adults, especially those aged 80 years and above, are particularly susceptible to a range of pathological factors that significantly influence their capacity to undergo and recover from LEAs, resulting in an increased post‐amputation mortality rate. 12 The research questions for the study were:
How was age group (aged ≥80 years and aged <80 years) associated with DFU outcomes in patients in the DLS programme?
How effective was the DLS programme in improving DFU outcomes in patients aged ≥80 years compared to historical controls?
We hypothesised that (1) age group was not associated with minor LEAs but was associated with higher risk of major LEAs, higher mortality and more healthcare utilisation in patients in the DLS programme and (2) the DLS programme had favourable impacts on DFU outcomes (e.g., lowered major LEA rate, increased LEA‐free survival rate and reduced healthcare utilisations) in patients aged ≥80 years.
2. MATERIALS AND METHODS
2.1. Study design and population
This observational cohort study encompassed three distinct patient groups: Group 1 consisted of elderly patients aged ≥80 years with DFUs who were enrolled in the DLS programme from June 2020 to July 2021; Group 2 comprised younger patients aged <80 years with DFUs who were enrolled in the DLS programme during the same timeframe and Group 3 constituted the historical control group, comprising elderly patients aged ≥80 years who were diagnosed with DFUs and received usual care between June 2016 and December 2017.
Patients in the DLS programme were recruited when they initially presented with DFUs at various healthcare settings where the DLS programme was implemented. The study settings included seven primary care polyclinics, two tertiary care clinics operating the Lower Extremity Amputation Prevention Programme (LEAPP) in two general hospitals and DFU‐related inpatient wards. Patients with DFUs were defined as those with DM and tissue loss, either anatomically at or distal to the malleolus, with conditions such as ulcer, abscess, osteomyelitis, cellulitis, with or without neuro‐osteoarthropathy. Patients with DM but with ulcers proximal to the malleoli (mixed arterio‐venous aetiology), callus only, cellulitis only, peripheral arterial disease without tissue loss and pre‐diabetes were excluded from the study.
Figure 1 illustrates the flow charts of patient selection in each group. Group 1 had 406 DLS programme patients aged ≥80 years (14.5% of the total DFU patients in the DLS programme). Group 2 consisted of 2392 DLS programme patients aged <80 years (31.5% of the total DFU patients in control cohort). Group 3 comprised 1718 DFU patients aged ≥80 years in the historical control group.
FIGURE 1.
Flow charts of patient selection in each. DFU, diabetic foot ulcer; DLS, diabetic limb salvage.
As the primary analysis of this study focused on estimating the average treatment effect (ATE) to assess the effectiveness of the MTD DLS programme in reducing major LEAs among DFU patients ≥80 years, we estimated the minimum sample size required for comparing two proportions. 19 Considering that the ATE reflected the difference in the probability of major LEAs between two groups—Group 1 (programme patients ≥80 years) with a one‐year major LEA rate of 3% and Group 3 (historical control patients ≥80 years) with a rate of 9% 18 and given an alpha level of 0.05, a power of 0.80 and a 1:1 ratio of sample sizes in the two groups, we calculated that a sample size of 381 patients per group would be required. Hence, the number of participants in our study was adequate to evaluate the major LEA rate as the primary outcome.
2.2. The DLS programme
The core elements of the multi‐disciplinary DLS programme encompassed an enhanced and expanded DM Foot Screening and Surveillance, Treatment, Escalation Programme for ulcer prevention (STEP) in seven polyclinics and two LEAPP clinics situated in two hospitals. The DM Foot STEP targeted early identification and treatment, and rapid referral of patients with DFUs, alongside close collaboration with LEAPP clinics. The existing LEAPP clinic services were administered by a multi‐disciplinary team (MDT) comprising vascular surgery, endocrinology (with diabetic nurse educators) and podiatry and wound care nurses, with regular inputs from orthopaedic surgery. Key interventions of the LEAPP clinics involve ensuring prompt access, optimising glycaemic control, addressing medical risk factors, initiating revascularisation, implementing active wound care, providing proper offloading and conducting patient education. 17 , 18 To further bolster patient care coordination, the programme deployed four diabetic foot coordinators to integrate and streamline care across polyclinics and hospitals, oversee patient adherence and monitor outcomes. Additionally, three new podiatry posts were added to enhance existing podiatry services and facilitate the integration of DM Foot STEP services with the LEAPP clinic services. In total, the MDT DLS programme engaged over 60 healthcare professionals spanning from primary to tertiary institutions.
2.3. Data source and retrieval
Data for all patients prospectively recruited to the programme were stored within the standing programme database. All patients' demographics, clinical, administrative and healthcare data were retrieved by a designated data officer from the healthcare cluster's chronic disease management registry for diabetes, 20 with relevant International Classification of Diseases 10th Revision diagnosis codes, surgical procedure codes and service codes. All data were de‐identified and assigned a unique record ID for each participant for data merging and analyses.
2.3.1. Outcome variables
The primary outcomes were incidence of minor and major LEA and all‐cause mortality within 1 year from the date of enrolment. A minor LEA refers to the amputation of the foot or parts of the foot while a major LEA involves the removal of any part of the lower limb proximal to the ankle joint, including below‐knee and above‐knee amputations. 21 To account for the impact of increased lifespan on the incidence of LEA, a composite outcome of LEA and mortality, LEA‐free survival rate, was also analysed. The secondary outcomes were number of admissions at inpatient wards in National Healthcare Group institutions and cumulative length of stay (LOS), day surgery (DS) encounters, emergency department (ED) and specialist outpatient clinic (SOC) visits and primary care visits at the polyclinics within 1 year from the date of enrolment. As the dates were provided in MMYYYY format, the first day of the respective month was used as the proxy date for indexing and cumulative LOS calculation.
2.3.2. Independent/exposure variable
The independent/exposure variable of interest was the age group of patients in the DLS programme (Group 1: Programme patients aged ≥80 years vs. Group 2: Programme patients aged <80 years) for the first research question and the participation in the DLS programme (Group 1: Programme patients aged ≥80 years vs. historical control with patients aged ≥80 years) for the second research question.
2.3.3. Covariates
The covariates encompassed baseline demographic and clinical factors. Demographic variables included age, gender and ethnicity (Chinese vs. non‐Chinese). Clinical variables comprised patients' medical history of prior LEAs in the past 3 years, hypertension, dyslipidaemia, ischemic heart disease (IHD), stroke and chronic kidney disease (CKD) severity; medications including anti‐diabetic, anti‐coagulant, anti‐hypertensive and lipid‐lowering medications. In addition, baseline haemoglobin A1c (HbA1c) levels, measured not more than 3 months before or after enrolment for DLS programme patients, or the first DFU‐related encounter date for patients in the historical control group, were also retrieved.
2.4. Statistical analysis
Baseline characteristics of patients in each group were described using frequencies (n) and percentages (%) for categorical variables and using means with standard deviations (SD) for continuous variables. The comparisons of baseline characteristics between Groups 1 and 2 and between Groups 1 and 3 were conducted using Chi‐square tests for categorical variables and using the independent t‐tests or Wilcoxon rank‐sum tests for continuous variables, depending on the distribution of the data.
The 1‐year primary outcomes (minor and major LEA rates, all‐cause mortality rate and LEA‐free survival rate) in each group were described using frequencies and percentages. The unadjusted differences in primary outcomes between Groups 1 and 2 and between Groups 1 and 3 were examined using the Chi‐square test. The one‐year secondary outcomes (number of inpatient admissions, cumulative LOS, DS utilisation, ED, SOC and polyclinic visits) were described using mean ± SD as well as median with the first (Q1) and third quartiles (Q3), and their differences between Groups 1 and 2 and between Groups 1 and 3 were compared using the Mann Whitney U test since they had skewed distributions.
Multiple logistic regressions were conducted to examine the association between age group (Groups 1 and 2, independent variable) and primary outcomes (dependent variables) and Poisson regressions were performed to determine the association between age group (Groups 1 and 2, independent variable) and healthcare utilisation in individual care settings (dependent variables), adjusting for gender, ethnicity, medical history of variable chronic conditions and different types of medications. Baseline HbA1c reading was not adjusted in the final models due to incomplete data. Results did not differ significantly with or without adjustment for HbA1c. Odds ratios (ORs) or incidence‐rate ratios (IRRs) and 95% confidence intervals (CIs) were reported.
To estimate the treatment effect of the DLS programme (Groups 1 and 3, exposure variable) on both primary and secondary outcomes, estimation of the ATEs based on propensity scores was conducted. Coefficients with 95% CIs were reported. For easier interpretation, the coefficients for binary outcomes were converted to ORs by taking the exponent in the tables. Since the coefficients for continuous outcomes are typically the average difference in the outcome variables between the programme and historical control groups, they were renamed to ‘Difference’ with values unchanged. As a non‐experimental causal inference technique, propensity score matching (PSM) using logistic regression was employed with the aim to balance programme and historical control groups on confounding factors (age, female, Chinese, history of IHD and stroke, CKD severity and LEA history in past 3 years), to make them comparable. The selection of covariates for inclusion in the PSM was guided by theoretical relevance and clinicians' perceptions on their associations with programme participation as well as data availability in the study. The overlap in the range of propensity scores between the programme and historical control groups was evaluated using a graph illustrating the distribution of propensity scores (termed ‘balance’) and was examined using the standardised difference and variance ratio metrics. The standardised difference was evaluated against a common threshold of 10% or less, while the variance ratio was assessed using a threshold within a range of 0.5–2.
2.5. Sensitivity analysis
We conducted a sensitivity analysis to examine the treatment effect of the DLS programme on outcomes by further including baseline HbA1c for PSM and adjustment, with patients with missing data for baseline HbA1c omitted from the regression models.
All data analyses were conducted using Stata/SE 17.1 for Windows (StataCorp, College Station, TX). A two‐tailed p value of 0.05 was set as the level of significance for all tests.
2.6. Ethical considerations
This study was approved by the institutional domain‐specific review board (Reference number: 2021/01151). As a study aiming to improve the quality of care using anonymised administrative data, no informed consent from patients was required.
3. RESULTS
3.1. Patient characteristics
The characteristics of the patients in the respective groups are presented in Table 1. Patients in Group 1 (DLS programme patients aged ≥80 years) had a mean age of 85.6 years, while Groups 2 and 3 had a mean age of 62.4 years and 86.1 years, respectively. Over 80% of the patients in Groups 1 and 3 were Chinese, and more than half were females. For Group 2 (DLS programme patients aged <80 years), 51.5% were Chinese and the majority were males (64.6%).
TABLE 1.
Comparison of baseline characteristics between Groups 1 and 2 and between Groups 1 and 3.
| Baseline characteristics | Group 1: programme patients ≥80 years (n = 406) | Group 2: programme patients <80 years (n = 2392) | Group 3: historical control patients ≥80 years (n = 1718) | p‐value (Group 1 vs. Group 2) | p‐value (Group 1 vs. Group 3) |
|---|---|---|---|---|---|
| Age, mean (SD) | 85.6 (4.7) | 62.4 (10.3) | 86.1 (4.7) | <0.001 | 0.024 |
| HbA1c, mean (SD) | 7.2 (1.5) n = 338 | 8.2 (2.1) n = 2029 | 7.1 (1.6) n = 1609 | <0.001 | 0.140 |
| Female, n (%) | 234 (57.6) | 846 (35.4) | 1027 (59.8) | <0.001 | 0.429 |
| Chinese, n (%) | 328 (80.8) | 1232 (51.5) | 1392 (81.0) | <0.001 | 0.913 |
| Hypertension | 365 (89.9) | 846 (35.4) | 1671 (97.3) | <0.001 | <0.001 |
| Dyslipidaemia | 347 (85.5) | 1912 (79.9) | 1600 (93.1) | 0.009 | <0.001 |
| Ischaemic heart disease | 130 (32.0) | 698 (29.2) | 689 (40.1) | 0.247 | 0.003 |
| Chronic kidney disease (CKD) severity | <0.001 | <0.001 | |||
| No CKD | 81 (20.0) | 746 (31.2) | 210 (12.2) | ||
| Mild | 72 (17.7) | 820 (34.3) | 352 (20.5) | ||
| Moderate | 170 (41.9) | 472 (19.7) | 684 (39.8) | ||
| Severe | 83 (20.4) | 354 (14.8) | 472 (27.5) | ||
| Stroke | 100 (24.6) | 387 (16.2) | 794 (46.2) | <0.001 | <0.001 |
| Lower extremity amputation in past 3 years | 19 (4.9) | 278 (11.6) | 21 (1.2) | <0.001 | <0.001 |
| Anti‐diabetic drug | 283 (69.7) | 2022 (84.5) | 1037 (60.4) | <0.001 | <0.001 |
| Antiplatelet drug | 266 (65.5) | 1410 (59.0) | 1025 (59.7) | 0.012 | 0.030 |
| Anticoagulant drug | 66 (16.3) | 376 (15.7) | 252 (14.7) | 0.784 | 0.420 |
| Lipid lowering drug | 316 (77.8) | 1888 (78.9) | 1237 (72.0) | 0.617 | 0.017 |
| Anti‐hypertensive drug | 289 (71.2) | 1679 (70.2) | 1366 (79.5) | 0.686 | <0.001 |
Compared to Group 2 patients, Group 1 patients had a higher proportion of females, Chinese individuals, having a history of hypertension, dyslipidaemia and stroke, and were on anti‐diabetic and antiplatelet drugs (Table 1). Notably, the proportion of patients with previous LEA in the past 3 years was lower in Group 1 (4.9%) compared to Group 2 (11.6%).
Compared to Group 3 patients, Group 1 patients were younger, with lower rates of hypertension, dyslipidaemia, IHD, and stroke and a higher rate of previous LEA in the past 3 years (Table 1). Additionally, a higher proportion of Group 1 patients were on medications such as anti‐diabetic, antiplatelet, anti‐hypertensive and lipid‐lowering drugs (p < 0.05).
3.2. Research question 1: The association between age group and 1‐year outcomes in patients in the DLS programme
In patients aged ≥80 years in the DLS programme (Group 1), the minor and major LEA rate was 7.1% and 2.2%, respectively, and the one‐year mortality rate and the LEA‐free survival rate was 34.7% and 59.9%, respectively. The utilisation in different care settings was shown in Table 2. In comparison to patients aged <80 years (Group 2), patients aged ≥80 years (Group 1) had lower minor and major LEA rates but a higher mortality rate in 1 year. Hence, the LEA‐free survival rate was also lower (Table 2). Moreover, Group 1 patients had fewer DS visits, but slightly more inpatient admissions, although their cumulative LOS was shorter (p < 0.05).
TABLE 2.
One‐year outcomes in two age groups (Groups 1 and 2) before adjustment.
| One‐year outcomes | Group 1: programme patients ≥80 years (n = 406) | Group 2: programme patients <80 years (n = 2392) | p‐value | ||
|---|---|---|---|---|---|
| n (%)/mean (SD) | Median (Q1–Q3) | n (%)/mean (SD) | Median (Q1–Q3) | ||
| Primary outcomes | |||||
| Minor LEA | 29 (7.1) | ‐ | 288 (12.0) | ‐ | 0.004 |
| Major LEA | 9 (2.2) | ‐ | 152 (6.4) | ‐ | 0.001 |
| Mortality | 141 (34.7) | ‐ | 258 (10.8) | ‐ | <0.001 |
| LEA‐free survival | 243 (59.9) | ‐ | 1805 (75.5) | ‐ | <0.001 |
| Secondary outcomes | |||||
| Number of inpatient admissions | 1.4 (1.6) | 1 (0–2) | 1.3 (1.8) | 1 (0–2) | 0.041 |
| Cumulative length of stay | 17.3 (27.8) | 7 (0–24) | 18.2 (36.5) | 3 (0–19) | 0.007 |
| Number of DS utilisation | 0.2 (0.7) | 0 (0–0) | 0.4 (1.4) | 0 (0–0) | <0.001 |
| Number of ED visits | 1.5 (1.6) | 1 (0–2) | 1.4 (2.0) | 1 (0–2) | 0.074 |
| Number of SOC visits | 5.8 (7.2) | 3 (0–9) | 5.6(6.9) | 3 (0–8) | 0.817 |
| Number of polyclinic visits | 2.9 (3.9) | 1 (0–5) | 3.0 (3.8) | 1 (0–5) | 0.318 |
Note: p‐values for primary outcomes were computed using Chi‐square test and p‐values for secondary outcomes were computed using Mann–Whitney U test.
Abbreviations: DS, day surgery; ED, emergency department; LEA, lower extremity amputation; Q1, quartile 1; Q3, quartile 3; SD, standard deviation; SOC, specialist outpatient clinic.
Multiple logistic regression results, as illustrated in Table 3, demonstrated that programme patients aged ≥80 years (Group 1) had 67% lower odds of major LEA (OR: 0.33, 95% CI: 0.15, 0.72) and 36% lower odds of LEA‐free survival (OR: 0.64, 95% CI: 0.47, 0.88). The Poisson regression results showed that they had shorter cumulative LOS (IRR: 0.45, 95% CI: 0.29, 0.69) than those aged <80 years but no substantial differences were observed in number of inpatient admission or utilisation in other care settings.
TABLE 3.
Associations between age group and 1‐year outcomes using multiple logistic or Poisson regressions.
| One‐year outcomes | OR/IRR a | 95% CI | p‐value |
|---|---|---|---|
| Primary outcomes | |||
| Minor LEA | 0.86 | 0.52, 1.44 | 0.571 |
| Major LEA | 0.33 | 0.15, 0.72 | 0.005 |
| Mortality | 1.39 | 0.94, 2.06 | 0.098 |
| LEA‐free survival | 0.64 | 0.47, 0.88 | 0.006 |
| Secondary outcomes | 1.01 | 0.89, 1.15 | 0.869 |
| Number of inpatient admissions | 0.90 | 0.74, 1.09 | 0.276 |
| Cumulative length of stay | 0.45 | 0.29, 0.69 | <0.001 |
| Number of DS utilisation | 1.02 | 0.9, 1.16 | 0.713 |
| Number of ED visits | 1.06 | 0.92, 1.23 | 0.433 |
| Number of SOC visits | 0.95 | 0.83, 1.08 | 0.420 |
| Number of polyclinic visits | 0.86 | 0.52, 1.44 | 0.571 |
Abbreviations: DS, day surgery; ED, emergency department; IRR, incidence rate ratio; LEA, lower extremity amputation; OR, odds ratio; SOC, specialist outpatient clinic.
Reference group: patients aged <80 years in diabetic limb salvage programme, adjusted for gender, Chinese, history of hypertension, dyslipidaemia, ischemic heart disease, stroke, CKD severity, LEA in the past 3 years, anti‐diabetic, anti‐platelet, anticoagulant, anti‐hypertensive and lipid lowering drugs.
3.3. Research question 2: Treatment effects of the DLS programme on 1‐year outcomes in DFU patients aged ≥80 years
When comparing the 1‐year outcomes between patients aged ≥80 years in DLS programme (Group 1) and those in the historical control (Group 3), it was observed that Group 1 patients had a higher incidence rate of minor LEA, yet a lower mortality rate, resulting in a higher LEA‐free survival rate as indicated in Table 4. Compared to Group 3, patients in Group 1 had fewer ED visits and inpatient admissions, shorter cumulative LOS, but more polyclinic and SOC visits (p < 0.001).
TABLE 4.
Comparison of 1‐year outcomes between Groups 1 and 3 before matching.
| One‐year outcomes | Group 1: programme patients ≥80 years (n = 406) | Group 3: historical control (n = 1718) | p‐value | ||
|---|---|---|---|---|---|
| n (%)/mean (SD) | Median (Q1–Q3) | n (%)/mean (SD) | Median (Q1–Q3) | ||
| Primary outcomes | |||||
| Minor LEA | 29 (7.1) | ‐ | 20 (1.2) | ‐ | <0.001 |
| Major LEA | 9 (2.2) | ‐ | 42 (2.4) | ‐ | 0.787 |
| Mortality | 141 (34.7) | ‐ | 777 (45.2) | ‐ | 0.017 |
| LEA‐free survival | 243 (59.8) | ‐ | 901 (52.4) | ‐ | 0.007 |
| Secondary outcomes | |||||
| Number of inpatient admissions | 1.4 (1.6) | 1 (0–2) | 2.4 (1.8) | 2 (1–3) | <0.001 |
| Cumulative length of stay | 17.3 (27.8) | 7 (0–24) | 26.2 (28.2) | 17 (9–34) | <0.001 |
| Number of DS utilisation | 0.2 (0.7) | 0 (0–0) | 0.1 (0.5) | 0 (0–0) | 0.110 |
| Number of ED visits | 1.5 (1.6) | 1 (0–2) | 2.5 (2.0) | 2 (1–3) | <0.001 |
| Number of SOC visits | 5.8 (7.2) | 3 (0–9) | 2.6 (4.2) | 0 (0–4) | <0.001 |
| Number of polyclinic visits | 2.9 (3.9) | 1 (0–5) | 1.4 (2.8) | 0 (0–2) | <0.001 |
Note: p‐values for primary outcomes were computed using Chi‐square test and p‐values for secondary outcomes were computed using Mann–Whitney U test.
Abbreviations: DS, day surgery; ED, emergency department; LEA, lower extremity amputation; Q1, quartile 1; Q3, quartile 3; SD, standard deviation; SOC, specialist outpatient clinic.
3.3.1. Treatment effects of the DLS programme
The PSM constructed using logistic regression showed that overlap assumption between the Groups 1 and 3 was satisfied (Figure 2) and all the selected covariates were balanced in the two groups after matching (Table 5).
FIGURE 2.
The estimated density of the predicted probabilities between Groups 1 and 3.
TABLE 5.
The standardised differences and variance ratios for the covariates before and after matching.
| Covariates used for matching | Standardised difference | Variance ratio | ||
|---|---|---|---|---|
| Before matching | After matching | Before matching | After matching | |
| Age | −0.12 | −0.01 | 1.01 | 1.11 |
| Female | −0.04 | 0.06 | 1.02 | 0.97 |
| Chinese | −0.01 | −0.03 | 1.01 | 1.05 |
| Ischemic heart disease | −0.17 | 0 | 0.91 | 1.00 |
| Chronic kidney disease severity | ||||
| Mild | −0.07 | −0.03 | 0.90 | 0.95 |
| Moderate | 0.04 | 0.06 | 1.02 | 1.02 |
| Severe | −0.17 | 0.02 | 0.82 | 1.02 |
| Stroke | −0.46 | −0.01 | 0.75 | 1.00 |
| LEA history in past 3 years | 0.21 | 0.03 | 3.70 | 1.24 |
Abbreviation: LEA, lower extremity amputation.
Table 6 presents the ATEs of the DLS programme on the primary and secondary outcomes as compared to the historical control group after PSM. Elderly patients in the DLS programme had 5% higher probability of minor LEA (coefficient: 0.05, 95% CI: 0.02, 0.09). In terms of healthcare utilisations, they had fewer inpatient admissions (coefficient: −0.90, 95% CI: −1.14, −0.67), shorter cumulative LOS (coefficient: −8.93, 95% CI: −12.11, −5.74), and fewer ED visits (coefficient: −0.89, 95% CI: −1.12, −0.66), but more SOC (coefficient: 3.57, 95% CI: 2.63, 4.52) and polyclinic visits (coefficient: 1.54, 95% CI: 1.06, 2.03).
TABLE 6.
The average treatment effects of the diabetic limb salvage programme on the primary and secondary outcomes (n = 2124).
| One‐year outcomes | Coefficient | 95% CI | Odds ratio/difference | 95% CI | p‐value |
|---|---|---|---|---|---|
| Primary outcomes | |||||
| Minor LEA | 0.05 | 0.02, 0.09 | 1.05 | 1.02, 1.09 | 0.005 |
| Major LEA | 0.003 | −0.02, 0.03 | 1.00 | 0.98, 1.03 | 0.753 |
| Mortality | −0.06 | −0.13, 0.002 | 0.94 | 0.88, 1.00 | 0.058 |
| LEA‐free survival | 0.05 | −0.03, 0.12 | 1.05 | 0.97, 1.13 | 0.207 |
| Secondary outcomes | |||||
| Number of inpatient admissions | −0.90 | −1.14, −0.67 | −0.90 | −1.14, −0.67 | <0.001 |
| Cumulative length of stay | −8.93 | −12.11, −5.74 | −8.93 | −12.11, −5.74 | <0.001 |
| Number of DS utilisation | 0.06 | −0.02, 0.14 | 0.06 | −0.02, 0.14 | 0.142 |
| Number of ED visits | −0.89 | −1.12, −0.66 | −0.89 | −1.12, −0.66 | <0.001 |
| Number of SOC visits | 3.57 | 2.63, 4.52 | 3.57 | 2.63, 4.52 | <0.001 |
| Number of polyclinic visits | 1.54 | 1.06, 2.03 | 1.54 | 1.06, 2.03 | <0.001 |
Note: Adjusted for age, gender, ethnicity, medical history of hypertension, dyslipidaemia, ischemic heart disease and stroke, baseline chronic kidney disease severity, LEA in the past 3 years, medications including anti‐diabetic, anti‐platelet, anti‐coagulant, anti‐hypertensive and lipid lowering drugs.
Abbreviations: DS, day surgery; ED, emergency department; LEA, lower extremity amputation; SOC, specialist outpatient clinic.
3.3.2. Sensitivity analysis
The results of the SA are summarised in Table 7. When baseline HbA1c was also included in the model, other than the significant differences in the outcomes mentioned above, elderly patients in the DLS programme also had 11% (95% CI: −0.17, −0.04) lower probability of 1‐year mortality and 9% (95% CI: 0.02, 0.16) higher probability of LEA‐free survival.
TABLE 7.
The average treatment effect of the diabetic limb salvage programme on the primary and secondary outcomes with baseline HbA1c included in the models (n = 1947).
| One‐year outcomes | Coefficient | 95% CI | Odds ratio/difference | 95% CI | p‐value |
|---|---|---|---|---|---|
| Primary outcomes | |||||
| Minor LEA | 0.05 | 0.02, 0.08 | 1.05 | 1.02, 1.08 | <0.001 |
| Major LEA | −0.004 | −0.02, 0.01 | 1.00 | 0.98, 1.01 | 0.681 |
| Mortality | −0.11 | −0.17, −0.04 | 0.90 | 0.84, 1.04 | 0.001 |
| LEA‐free survival | 0.09 | 0.02, 0.16 | 1.09 | 1.02, 1.17 | 0.010 |
| Secondary outcomes | |||||
| Number of inpatient admissions | −0.82 | −1.08, −0.57 | −0.82 | −1.08, −0.57 | <0.001 |
| Cumulative length of stay | −9.50 | −13.03, −5.97 | −9.50 | −13.03, −5.97 | <0.001 |
| Number of DS utilisation | 0.12 | −0.02, 0.26 | 0.12 | −0.02, 0.26 | 0.089 |
| Number of ED visits | −0.76 | −1.03, −0.50 | −0.76 | −1.03, −0.50 | <0.001 |
| Number of SOC visits | 4.11 | 3.11, 5.11 | 4.11 | 3.11, 5.11 | <0.001 |
| Number of polyclinic visits | 1.84 | 1.35, 2.35 | 1.84 | 1.35, 2.35 | <0.001 |
Note: Adjusted for age, gender, ethnicity, medical history of hypertension, dyslipidaemia, ischemic heart disease and stroke, baseline chronic kidney disease severity, LEA in the past 3 years, baseline HbA1c, medications including anti‐diabetic, anti‐platelet, anti‐coagulant, anti‐hypertensive, and lipid lowering drugs.
Abbreviations: DS, day surgery; ED, emergency department; LEA, lower extremity amputation; SOC, specialist outpatient clinic.
4. DISCUSSION
This study delved into a comparative analysis of the impact of the MTD DLS programme on one‐year outcomes between patients aged ≥80 years and those aged <80 years and examined the treatment effects of the DLS programme on these outcomes as compared to matched historical controls. As there exists a paucity of literature focusing on the outcomes of multi‐disciplinary diabetic foot care programmes in DFU patients aged ≥80 years, this study contributes valuable insights and evidence to a relatively less explored area of research.
Several studies have explored the complex interplay between age and the efficacy of DLS strategies, yielding contradictory findings in the existing literature. While some literature suggested that advanced age might be associated with poorer outcomes, 22 , 23 others demonstrated that limb salvage could be achieved in older DFU patients at rates comparable to their younger counterparts, despite these patients presenting with more comorbidities and foot‐related complications. 24 , 25 , 26 The findings of this study showed that compared to patients <80 years, those aged ≥80 years had lower adjusted odds of major LEA and LEA‐free survival, which is potentially attributable to the higher mortality rate observed in patients aged ≥80 years. While there was no significant difference in number of inpatient admissions between the age groups, patients aged ≥80 years exhibited a shorter LOS, potentially indicative of the DLS programme's success in averting DFU complications and improving patients' overall health. This shorter LOS may be partially attributed to the inclusion of a specialised geriatric surgical service 27 , 28 within the inpatient diabetic foot pathway. These findings resonate with a previous study demonstrating improved limb salvage outcomes through revascularisation in patients over 80 years old with critical limb ischemia. 25 However, as the study did not manage to recruit all DFU patients in the DLS programme and did not collect data on clinicians' assessments of the necessity of LEA for individual patients and their decision‐makers' stance on the procedure, we could not conclude that the DLS programme is more effective in preventing major LEA in patients aged ≥80 years as we could not rule out the possibility that the LEA recommendation rate was lower or the LEA refusal rate was higher in patients aged ≥80 years than that in patients <80 years.
When evaluating the treatment effects of the DLS programme in the elderly population, the favourable results such as a relatively lower mortality rate, a higher LEA‐free survival rate, reduced ED visits and inpatient admissions, and shorter LOS provide encouraging evidence. These outcomes collectively suggest that the multi‐disciplinary DFU care management approach implemented through the DLS programme has yielded promising effects on patient outcomes, although the probability of minor LEA was 5% higher in the elderly patients in the programme. In addition, the increased utilisation of SOC and polyclinic services indicates a noteworthy shift of care from costly inpatient tertiary care to primary and secondary care settings, aligning with the programme's objectives. The findings echo other studies that have demonstrated the favourable impact of the multi‐disciplinary management approach for DFUs, both in terms of clinical and economic outcomes. 18 , 29 , 30 However, it is important to exercise caution when attributing the observed differences solely to the DLS programme as comparing outcomes between a current programme group and the historical control group does not account for changes in medical practices, technological advancement, and patient management which may have evolved over time.
This study has several limitations. Firstly, we did not collect on some potential confounding factors, for example, patients' socioeconomic factors, the duration of diabetes, DFU characteristics and complications such as skin infections, abscess formation and gangrene, baseline Charlson Comorbidity Index and frailty state and adherence to medication and programme. These unobserved data may deviate the results or the effectiveness of the programme. Secondly, the use of a historical control group, due to the lack of randomisation, may introduce selection bias as patients in the historical control group might differ from those in the programme, potentially leading to biased comparisons. 31 While PSM was applied to mitigate imbalances between the programme and historical control groups, it only addressed observed characteristics, leaving the potential for bias from unobserved variables that could affect DFU outcomes and thus the accuracy of the ATE estimates. 32 Thirdly, some temporal shifts such as changes in standard care practices, diagnostic criteria and overall healthcare quality over time could independently affect the programme effect, making it challenging to attribute the observed differences solely to the programme. Fourthly, we did not account for patients' compliance to the protocol of the DLS programme in our analyses, which could have impacted the estimated effect size of the programme. Despite these limitations, utilising a historical control to adjust or match for key confounders including demographics, history of chronic medical conditions and common medications offers valuable insights. Finally, as the study only sampled patients who had been enrolled in the DLS programme for at least 1 year, the findings may not be generalisable to all programme participants.
5. CONCLUSION
In conclusion, our study suggests that the MDT DLS programme may offer beneficial outcomes for elderly patients aged ≥80 years with DFUs as it demonstrated reduced likelihood of major LEA rates and an improved LEA‐free survival rate compared to DFU patients aged <80 years. However, the findings should be interpreted with caution due to the inherent limitations of using historical controls. Their generalisability to other elderly populations in different healthcare contexts or with varying demographic profiles might be limited. The positive outcomes observed underline the need for targeted interventions for DFUs in the elderly, considering their unique biological and social conditions, as well as specific needs such as multimorbidity, frailty state and advance care planning. Future research is essential to validate our findings in a broader and more diverse population of elderly patients with proper selection of concurrent controls and to further explore the underlying mechanisms driving the observed differences in outcomes.
FUNDING INFORMATION
The programme receives funding support offered by the Skin Research Institute of Singapore, Agency for Science, Technology and Research (A*STAR) under its Industry Alignment Fund—Pre‐Positioning Programme (IAF‐PP) as part of Wound Care Innovation for the Tropics (WCIT) Programme (Grant Number H17/01/a0/0Y9), the National Healthcare Group's Population Health Grant (Grant Number PHG20/S/X/1/1), and the National Medical Research Council Grant (Grant / Award Number: FLWSHP19nov‐0015).
CONFLICT OF INTEREST STATEMENT
The authors declare that there is no financial relationship and conflict of interest.
Ge L, Zhao J, Tan M, et al. Multi‐disciplinary diabetic limb salvage programme in octogenarians with diabetic foot ulcers is not futile: An observational study with historical controls. Int Wound J. 2024;21(3):e14801. doi: 10.1111/iwj.14801
DATA AVAILABILITY STATEMENT
According to the Data Protection Act Commission Singapore—Advisory Guidelines for the Healthcare Sector, the personal health data collected for the study are not publicly available due to legal and ethical restrictions related to data privacy protection. However, the minimal dataset underlying the findings in the manuscript is available upon request to interested researchers after authorisation of the ethical committee of the National Health Group Domain Specific Review Board. Interested researchers may contact Dr Joseph Lo (zhiwen_lo@wh.com.sg) for requests for data.
REFERENCES
- 1. Saeedi P, Petersohn I, Salpea P, et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: results from the international diabetes federation diabetes atlas, 9th edition. Diabetes Res Clin Pract. 2019;157:107843. [DOI] [PubMed] [Google Scholar]
- 2. Ministry of Health Singapore . Disease burden. Accessed August 15, 2023. https://www.moh.gov.sg/resources-statistics/singapore-health-facts/disease-burden
- 3. Phan TP, Alkema L, Tai ES, et al. Forecasting the burden of type 2 diabetes in Singapore using a demographic epidemiological model of Singapore. BMJ Open Diabetes Res Care. 2014;2(1):e000012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Li M. Guidelines and standards for comprehensive clinical diagnosis and interventional treatment for diabetic foot in China (issue 7.0). J Interven Med. 2021;4(3):117‐129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Bellia A, Meloni M, Andreadi A, Uccioli L, Lauro D. Geographic and ethnic inequalities in diabetes‐related amputations. Front Clin Diab Healthcare. 2022; 3:855168. doi: 10.3389/fcdhc.2022.855168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Riandini T, Pang D, Toh MPHS, et al. National rates of lower extremity amputation in people with and without diabetes in a multi‐ethnic Asian population: a ten year study in Singapore. Eur J Vasc Endovasc Surg. 2022;63(1):147‐155. [DOI] [PubMed] [Google Scholar]
- 7. Ezzatvar Y, García‐Hermoso A. Global estimates of diabetes‐related amputations incidence in 2010–2020: a systematic review and meta‐analysis. Diabetes Res Clin Pract. 2023;195:110194. [DOI] [PubMed] [Google Scholar]
- 8. Soyoye DO, Abiodun OO, Ikem RT, Kolawole BA, Akintomide AO. Diabetes and peripheral artery disease: a review. World J Diabetes. 2021;12(6):827‐838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Chin SF. The Straits Times. Singapore's population ageing rapidly: nearly 1 in 5 citizens is 65 years and older. 2022. Accessed September 13, 2023. https://www.straitstimes.com/singapore/singapores-population-ageing-rapidly-184-of-citizens-are-65-years-and-older
- 10. Singapore Department of Statistics . Complete life tables for Singapore resident population 2021–2022. 2023. Accessed September 12, 2023. https://www.singstat.gov.sg/‐/media/files/publications/population/lifetable21‐22.ashx#:~:text=Life%20expectancy%20at%20birth%20of,declines%20in%202021%20and%202022
- 11. Leon LR, Psalms SB, Walters J. Diabetic foot infections in the elderly: primary amputation versus ‘foot‐sparing surgery’. A Case Report Int Wound J. 2007;4(4):315‐321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Klaphake S, de Leur K, Mulder PG, et al. Mortality after major amputation in elderly patients with critical limb ischemia. CIA. 2017;22(12):1985‐1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Schaper NC, van Netten JJ, Apelqvist J, et al. Practical guidelines on the prevention and management of diabetic foot disease (IWGDF 2019 update). Diabetes Metab Res Rev. 2020;36(Suppl. 1):e3266. [DOI] [PubMed] [Google Scholar]
- 14. Hingorani A, LaMuraglia GM, Henke P, et al. The management of diabetic foot: a clinical practice guideline by the Society for Vascular Surgery in collaboration with the American Podiatric Medical Association and the Society for Vascular Medicine. J Vasc Surg. 2016;63(2 Suppl):3S‐21S. [DOI] [PubMed] [Google Scholar]
- 15. Ibrahim A. IDF clinical practice recommendation on the diabetic foot: a guide for healthcare professionals. Diabetes Res Clin Pract. 2017;1(127):285‐287. [DOI] [PubMed] [Google Scholar]
- 16. Gambardella GV, Blume PA. Diabetic limb salvage: Surgeon's perspective. Lower Extremity Review Magazine. 2013. Accessed September 12, 2023. https://lermagazine.com/article/diabetic-limb-salvage-surgeons-perspective
- 17. Lo ZJ, Tan E, Chandrasekar S, et al. Diabetic foot in primary and tertiary (DEFINITE) care: a health services innovation in coordination of diabetic foot ulcer (DFU) care within a healthcare cluster—18‐month results from an observational population health cohort study. Int Wound J. 2022;20(5):1609‐1621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Lo ZJ, Chandrasekar S, Yong E, et al. Clinical and economic outcomes of a multidisciplinary team approach in a lower extremity amputation prevention programme for diabetic foot ulcer care in an Asian population: a case‐control study. Int Wound J. 2022;19(4):765‐773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions. J. Wiley; 2003:79. [Google Scholar]
- 20. Heng BH, Sun Y, Cheah JTS, Jong M. The Singapore National Healthcare Group Diabetes Registry—Descriptive epidemiology of type 2 diabetes mellitus. Ann Acad Med Singapore. 2010;39(5):348‐352. [PubMed] [Google Scholar]
- 21. Pran L, Harnanan D, Baijoo S, et al. Major lower limb amputations: recognizing pitfalls. Cureus. 2021;13(8):e16972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Primadhi RA, Septrina R, Hapsari P, Kusumawati M. Amputation in diabetic foot ulcer: a treatment dilemma. World J Orthop. 2023;14(5):312‐318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Morbach S, Furchert H, Gröblinghoff U, et al. Long‐term prognosis of diabetic foot patients and their limbs: amputation and death over the course of a decade. Diabetes Care. 2012;35(10):2021‐2027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Aragón‐Sánchez J, Víquez‐Molina G, López‐Valverde ME, Aragón‐Hernández C, Aragón‐Hernández J, Rojas‐Bonilla JM. Clinical features, inflammatory markers, and limb salvage in older adults with diabetes‐related foot infections. Int J Low Extrem Wounds. 2023; 15347346231154472. doi: 10.1177/15347346231154472. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 25. Dosluoglu HH, Lall P, Cherr GS, Harris LM, Dryjski ML. Superior limb salvage with endovascular therapy in octogenarians with critical limb ischemia. J Vasc Surg. 2009;50(2):305‐316.e2. [DOI] [PubMed] [Google Scholar]
- 26. Choi MSS, Jeon SB, Lee JH. Predictive factors for successful limb salvage surgery in diabetic foot patients. BMC Surg. 2014;14(1):113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Natesan S, Li JY, Kyaw KK, et al. Effectiveness of comanagement model: geriatric medicine and vascular surgery. J Am Med Dir Assoc. 2022;23(4):666‐670. [DOI] [PubMed] [Google Scholar]
- 28. Yong E, Gong H, Liew H, et al. Getting a foothold on diabetic foot disease‐outcomes of a multidisciplinary clinical pathway for inpatient diabetic foot care: a 17‐year institutional review. Int J Low Extrem Wounds. 2023;15347346231183740. doi: 10.1177/15347346231183740. Online ahead of print. [DOI] [PubMed] [Google Scholar]
- 29. Wang C, Mai L, Yang C, et al. Reducing major lower extremity amputations after the introduction of a multidisciplinary team in patient with diabetes foot ulcer. BMC Endocr Disord. 2016;16:38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Almdal T, Nielsen AA, Nielsen KE, et al. Increased healing in diabetic toe ulcers in a multidisciplinary foot clinic‐an observational cohort study. Diabetes Res Clin Pract. 2015;110(3):315‐321. [DOI] [PubMed] [Google Scholar]
- 31. Viele K, Berry S, Neuenschwander B, et al. Use of historical control data for assessing treatment effects in clinical trials. Pharm Stat. 2014;13(1):41‐54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Clarke GM, Conti S, Wolters AT, Steventon A. Evaluating the impact of healthcare interventions using routine data. BMJ. 2019;365:l2239. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
According to the Data Protection Act Commission Singapore—Advisory Guidelines for the Healthcare Sector, the personal health data collected for the study are not publicly available due to legal and ethical restrictions related to data privacy protection. However, the minimal dataset underlying the findings in the manuscript is available upon request to interested researchers after authorisation of the ethical committee of the National Health Group Domain Specific Review Board. Interested researchers may contact Dr Joseph Lo (zhiwen_lo@wh.com.sg) for requests for data.