Importance. Conventionally, treatment goals for diabetic patients primarily target glycemic levels and traditional cardiovascular risk factors (blood pressure, lipids) control to reduce macro- and micro-vascular complications. More recently, the relevance of assessing functional status in older diabetic patients has emerged. A knowledge gap exists regarding the risk of functional dependency in older diabetic patients and on the impact of Comprehensive Geriatric Assessment (CGA) on the achievement of easily calculated and objective patient-centered outcomes.

Objective To investigate if a personalized plan of care base on CGA would reduce the risk of functional decline over time in older patients with diabetes as compared to usual care Design. Individual patient randomized controlled trial comparing intervention with usual care.

Setting. Three hospitals in Cagliari, Ferrara, and Milano, Italy. Participants. One-hundred and eighty diabetic patients aged ≥75 years. Intervention. Usual care for both the Control and Intervention groups will be assured by a diabetologist according to in-use guidelines. After randomization, a geriatrician will administer a thorough CGA to all participants allocated to the intervention groups. CGA will be used to identify specific treatment goals according to the functional status of the patient and to incorporate Patient-preferred outcome in the management of diabetes and comorbidities. Using the results of the CGA the geriatricians along with the attending diabetologist and additional healthcare professionals, if needed, will implement a personalized diagnostic and therapeutic plan of care. Participants in the Control group will receive no additional intervention over and above usual care.

Follow-up. Follow-up visit will be scheduled at 6 and 12 months after randomization.

Main outcome measures. The primary outcome will be represented by the change in physical performance, assessed by change in the Short Physical Performance Battery (SPPB) score over time.

Secondary outcomes will be represented by change in cognitive function, sarcopenia, dependency, glycated hemoglobin levels, and rate of hospitalization as well as Time-at-home.

Expected Results. The CGA-driven intervention applied to older diabetic patients will have significant benefits on functional outcomes as compared to usual care.

Implications. Multimodal intervention in older diabetic patients will significantly impact on the ageing population and allow a novel process to be developed for interventions that produce the maximum disability-free life years lived combined with the highest quality of life for this vulnerable and often neglected group of adults.

Diabetes is highly prevalent in older subjects, affecting more than 25% of people over the age of 65. Screening, intervention, and monitoring of older subjects with diabetes requires facing complexity. Older diabetic patients are at greater risk of macro-vascular cardiovascular (CV) complications (myocardial infarction, stroke, peripheral arterial disease). Because of micro-vascular complications of diabetes (falls because of neuropathy, walking limitation because of peripheral artery disease or leg amputation, blindness, etc.), coexisting chronic disease, and consequent polypharmacy, older subjects with diabetes are at dramatically greater risk of disability and functional loss. Older diabetic patients have higher rates of premature death (4.6 years earlier), functional disability, accelerated muscle loss, and develop disability 6 to 7 years earlier than peers without diabetes. Too often, in routine clinical practice, the decision-making process is characterized by a fundamental dilemma: the older diabetic patient for whom the clinician is recommending treatment according to the Guidelines is quite different from the inclusion criteria adopted in the studies the guidelines are based on. Last, but not least, older patients often underestimate their risk of adverse outcomes. Care incorporating patient preferred/reported outcome may facilitate communication concerning the clinical course of diabetes and can increase the attention and participation of patients and caregivers. Indeed, a growing proportion of older diabetic subjects give higher importance to outcomes such as functional independence and related quality of life than on short-term complications or survival. Not to mention that also the preference for a specific functional outcome differs from patient to patient.

In this context, the conventional approach to the diabetic patient - focusing primarily on metabolic and CV risk factor control - does not meet the needs of older diabetic subjects and of their caregivers. Comprehensive Geriatric Assessment (CGA) is an effective tool for the holistic evaluation of the older patients through a comprehensive assessment of the physical, psychological, functional and socio-economic dimensions. Thus, CGA is intended to guide an overall approach to care from the determination of targets and therapeutic approaches to identification of patient's ability for diabetes self-management, from the aspect of patient's preferences and needs to the improvement or maintenance of functions and residual ability. However, CGA in older diabetic subjects is still predominantly adopted only to determine targets for glucose and CV risk factor levels A knowledge gap exists regarding the risk of functional dependency in older diabetic patients and the impact of CGA on the achievement of easily calculated and objective patient-centered outcomes.

METHODOLOGY Study design This project consists of a multicenter parallel groups randomized controlled trial conducted in three diabetological outpatient clinics in the University Centers in Cagliari, Ferrara, and Milan, Italy.

Participants Patients will be eligible if affected by diabetes and aged ≥75 years. Inclusion criteria: Age ≥75 years; diagnosis of diabetes mellitus; functional limitation defined as a Short Physical Performance Battery (SPPB) score between 4 (included) and 9 (included), and willingness to participate in the study. Exclusion criteria: severe disability in basic activity of daily living (dependency in 3 or more activities), not being resident in the hospital catchment area; residence in long-term care; diagnosis of schizophrenia, other psychotic or bipolar disorder, or severe cognitive impairment (MMSE score < 21/30); alcohol abuse (> 14 drinks per week); cancer requiring treatment in the past 3 years, except for non-melanoma skin cancers or cancers that have an excellent prognosis (e.g., the early stage breast or prostate cancer); respiratory insufficiency requiring regular use of supplemental oxygen; recent (< 6 months) myocardial infarction; class IV NYHA heart failure; severe chronic kidney disease (stage V, dialysis); decompensated liver cirrhosis, inability or unwillingness to provide informed consent. Recruitment All patients attending the diabetological outpatient clinics will be screened for potential participation in the study. Specifically trained researchers embedded in the medical units will assess participants for inclusion and exclusion criteria. Potentially eligible participants who agreed to participate will be finally evaluated with the Short Physical Performance Battery for final inclusion in the study. Patients with a SPPB score between 4 and 9, after giving written consent will be randomized to intervention or control group with a 1:1 ratio.

Randomisation. A secure internet-based system to randomize individual participants one to one between the intervention and control groups, with stratification by center, will be used.

Baseline Assessment At baseline both intervention and control group will be assessed according to usual diabetological evaluation procedures. Demographics, clinicals and diabetes-related information will be recorded. In addition, intervention and control groups will be assessed by CGA and measurements of muscle strength.

Interventions Usual care for both the Control and Intervention groups will be assured by a specialist physician in diabetology, according to in-use guidelines. After randomization, a dedicated researcher, specialist in geriatric medicine, will administer a thorough Comprehensive Geriatric Assessment to all participants allocated to the intervention groups. CGA results will be used to diagnoses or drug review; to identify specific treatment goals according to the functional status of the patient and to incorporate Patient-preferred outcome in the management of diabetes and comorbidities. Using the results of the CGA the geriatricians along with the attending diabetologist and additional healthcare professionals, if needed, will implement a personalized diagnostic and therapeutic plan of care. Participants in the Control group will receive no additional intervention over and above usual care.

Follow-up. Follow-up visit will be scheduled at 6 and 12 months after randomization. Follow-up visit will be performed at the out-patient clinic. For patients not able to travel to the out-patient clinics home visits will be provided. For institutionalized or hospitalized patients, a phone interview with relatives or caregivers will be organized.

Outcomes Primary outcome. The primary outcome will be represented by the change in physical performance, assessed by the Short Physical Performance Battery (SPPB) score over time.

Sample size and statistical analysis Sample size Using data from a previous observational study in older Italian people with diabetes (9), we hypothesized a baseline mean SPPB score of 6.9 with a 1.8 standard deviation. Assuming a 1 SPPB score point as a clinically meaningful SPPB difference (10) at the end of the follow-up between intervention and control groups, we estimated that 72 patients for each group would provide a 90% power at a 5% significance level to detect the one-point difference between the two groups. Assuming a 20% loss to follow-up, including patients who will die during the study follow-up, we are planning to enroll 90 patients for each study group for a total or 180 participants.

Statistical analysis Baseline characteristics will be compared between intervention and control groups using ANOVA for continuous normally distributed variables and chi-squared test for categorical variables. Non-parametric tests will be used for continuous variables with asymmetric distribution. Difference between study groups in SPPB score over the follow-up (primary outcome) will be estimated and compared using longitudinal data analysis for repeated measures techniques, adjusting for potential confounders. Incidence rate in secondary binary endpoints will be compared using multivariable logistic regression models or Cox proportional hazard model as appropriate. Change over time in secondary continuous outcomes will be compared across groups using longitudinal data analysis for repeated measures techniques.

OBJECTIVE Conventionally, treatment goals for diabetic patients primarily target glycemic levels and traditional cardiovascular risk factors (blood pressure, lipids) control to reduce macro- and micro-vascular diabetes complications. More recently, the relevance of assessing functional levels in older diabetic patients has emerged. We propose that functional outcomes should be adopted as treatment goals in older diabetic subjects to prevent, reduce or postpone functional decline. In other terms, performance measures offer relevant outcomes to act on the healthcare needs of older diabetic patients, to improve patients' quality of life and decrease the risk of functional dependency. As performance measures will become part of the usual care of older diabetic patients, clinicians will need easy access to such criteria to determine whether a change in performance in an individual patient is meaningful. For this purpose, we adopted functional outcomes that are easily and quickly measurable in a standardized and reliable manner in clinical practice. Furthermore, the primary outcome selected for this study, has been associated with a number on important clinical endpoint in both community-dwelling and hospitalized patients, including long-term disability, hospitalization, institutionalization, and death. Notably, the functional outcomes adopted by the present Project belong to the so-called Patient Reported Outcome Measures (PROMs). Therefore, their adoption as primary treatment goal in older diabetic patients will increase the engagement of patients and caregivers in the care process, and in fostering a more personalized, inclusive and participative medicine.

EXPECTED RESULTS The present Project will contribute to cumulate evidence concerning the impact of GCA and relevance of functional outcomes as key components of the process of care and therapeutic goals in older diabetic patients. The adoption and inclusion of performance measures may offer a powerful mechanism to understand and act on the healthcare needs of older adults and to implement the role of patients as the key Stakeholder of the healthcare system. Accordingly, Patient Reported Experience Measures (PREMs) and Patient Reported Outcome Measures (PROMs) will receive greater attention in the path of care for older patients. We expect that multimodal approach will help to preserve better motor function over time (primary goal of the present study) in diabetic patients than usual care. This will reinforce the routine adoption of a Comprehensive Geriatric Assessment in the routine clinical management of older diabetic patients and an integrated multidimensional and multiprofessional approach to individual care plans. We also expect that the multimodal approach will result in maintenance of multiple functions (cognition, muscular, etc.) with greater satisfaction of patients and caregivers and a greater time spent at home. In the long-term, these results will imply a considerable reduction in functional decline and progression to disability of older diabetic patients, with great saving for caregivers and the Public Healthcare System. The relatively short duration of follow-up allowed by the duration of the present Call will make the present study a proof-of-concept also concerning feasibility of such an innovative path of care. Accordingly, we will continue following-up patients enrolled in the present study for at least 12 months.

The present randomized study will contribute to closing the existing gap in knowledge about the usefulness of the standardized use of CGA and the adoption of functional outcomes as the primary goal for treatment of older diabetic subjects.

The adoption of functional outcomes in the treatment of older diabetic patients will also allow a better risk stratification in the initial assessment of patients, facilitating the predictive capacity of available risk scores, also with respect to more conventional outcomes as hospitalizations. Additionally, the proposed method to estimate the risk of adverse clinical events (including functional outcomes) may be able to better identify patients who will withdraw intervention because of poor effectiveness and/or iatrogenic side effects (by itself leading to hospitalizations and, eventually, to disability) thus resulting in savings for the National Health System

Inclusion criteria:

• Age ≥75 years;
• Diagnosis of diabetes mellitus;
• Functional limitation defined as a Short Physical Performance Battery (SPPB) score between 4 (included) and 9 (included), and
• Willingness to participate in the study.

Exclusion criteria:

• Severe disability in basic activity of daily living (dependency in 3 or more activities),
• Not being resident in the hospital catchment area; residence in long-term care;
• Diagnosis of schizophrenia, other psychotic or bipolar disorder, or severe cognitive impairment (MMSE score < 21/30);
• Alcohol abuse (> 14 drinks per week);
• Cancer requiring treatment in the past 3 years, except for non-melanoma skin cancers or cancers that have an excellent prognosis (e.g., the early stage breast or prostate cancer);
• Respiratory insufficiency requiring regular use of supplemental oxygen;
• Recent (< 6 months) myocardial infarction;
• Class IV NYHA heart failure;
• Severe chronic kidney disease (stage V, dialysis);
• Decompensated liver cirrhosis;
• Inability or unwillingness to provide informed consent.