Study design and participants
The ProGERO is a prospective cohort study of community-dwelling older adults from an outpatient clinic at the Hospital das Clinicas of the University of Sao Paulo Medical School (HCFMUSP), in Sao Paulo, Brazil. The study aims to explore sociodemographic and clinical characteristics associated with adverse outcomes during the study follow-up, including falls, disability, emergency room (ER) visits, hospital admissions, institutionalization, and death.
HCFMUSP is the largest academic medical center in Latin America, following 1.5 million persons (28% of them are older patients) from the metropolitan area of Sao Paulo, Brazil. HCFMUSP geriatrics clinic operates 12 hours a day, five days a week, with a multidisciplinary team (geriatricians, registered nurses, social workers, and psychologists) offering regular appointments (usually every three months) for older outpatients.
In our recruitment, we invited every patient aged 60 years and over who had a medical appointment at the clinic between April and December 2017. We excluded subjects according to the following criteria: (1) need for immediate hospital admission or emergency care on baseline (e.g., hemodynamic instability, acute respiratory symptoms, delirium); (2) inability to be reached by telephone for follow-up assessments between visits; or (3) refusal to consent with the study.
Eligible patients who consented to participate underwent a baseline clinical assessment with a standardized interview and physical examination. After recruitment, baseline characteristics will be reassessed every three years during in-person follow-up visits. We plan to invite new patients to participate in the study in each new wave of in-person visits and estimate to include approximately 700 new participants per recruitment cycle. We will also complete 6-month telephone interviews to collect data on our outcomes of interest between visits. Finally, we plan to follow participants for at least ten years, or until their deaths. The study design is further detailed in Fig. 1.
A multidisciplinary team of four registered nurses and four geriatricians completed the baseline clinical assessments. The nurses were responsible for undertaking the interviews and questionnaires, while geriatricians monitored data quality (review of missing data and information reliability), elucidated queries during assessments, and reviewed electronic medical records to collect data on multimorbidity and medications. When participants were unable to communicate, we interviewed family members and caregivers to obtain the best information available.
We collected and managed the data using the Research Electronic Data Capture (REDCap) .
Demographics. We collected the following sociodemographic data: age; sex; race/ethnicity; marital status; level of literacy; occupation; annual household income per capita, expressed both as a continuous variable and as categories according to the Brazilian minimum wage in 2017 [1 minimal wage = 4000 United States dollar (USD) per year]; and neighborhood. We also recorded whether participants lived alone or with other persons.
Multimorbidity and medications. We measured multimorbidity using the Charlson comorbidity index  and the Functional Comorbidity Index (FCI) , based on information retrieved from medical records. We also used medical records to compile the lists of medications in use.
The Charlson comorbidity index includes 19 clinical conditions, with various scoring weights, and the final score is defined by the total sum of items (range: 0–37 points; 37 = worst) . We analyzed the Charlson comorbidity index both as a continuous variable and stratified in ordinal categories (0, 1–2, and ≥ 3 points) . The FCI is a comorbidity scale designed to predict functional decline. It includes 18 clinical conditions, and its score corresponds to the total disease count (range: 0–18; 18 = worst) .
Anthropometry, physical examination, and sensory evaluation. Anthropometric and physical examination measures included: blood pressure, pulse rate, weight, height, and calf circumference.
Blood pressure and pulse rate were measured at the heart level, using an electronic manometer and a standardized cuff (Omron Hem-7113 Automatic Blood Pressure Monitor, Omron Healthcare Co., Ltd.). Values were recorded after 5 minutes of rest while sitting. Three readings were taken in succession, with at least one-minute intervals, and the average was used for the analyses .
We asked participants to wear light clothing and no shoes when anthropometric measures were taken. We calculated body mass index (BMI) using the metric system (kg/m2) and measured the calf circumference (cm) using an inelastic tape placed on the broadest possible section of the left calf .
Finally, we screened for the presence of visual and auditive deficits ("yes" or "no") using the following questions, extracted from the Alzheimer's Disease Cooperative Study - Activities of Daily Living - Prevention Instrument (ADCS-ADL-PI) Questionnaire : (1) "Can you see well enough to recognize a friend across the street?"; and (2) "Can you usually hear and understand another person when they talk in a normal voice?".
Comprehensive Geriatric Assessment. The 10-minute Targeted Geriatric Assessment (10-TaGA) is a validated multi-domain hands-on instrument that was developed to screen geriatric syndromes and estimate the global impairment of patients, using the
cumulative deficit model. In previous research, 10-TaGA provided adequate
validity and good accuracy in discriminating between frail and
non-frail individuals and good predictive power for one-year mortality, disability and hospitalization[14, 16, 17]. The 10-TaGA evaluates: (1) social support (living arrangements and availability of help) ; (2) emergency department visits and hospitalizations in the previous six months; (3) the number of falls in previous 12 months; (4) the number of medications; (5) dependence in activities of daily living (ADLs) (Katz index); (6) 10-point Cognitive Screener (10-CS) ; (7) self-rated health; (8) 4-item Geriatric Depression Scale (GDS-4) score ; (9) nutritional status (weight loss in the previous 12 months, BMI and calf circumference); (10) gait speed . Each domain is categorized and scored as normal (0 points), mild impairment (0.5 points), or severe impairment (1 point), based on conventional cut-off points derived from the literature for singular items. A single numerical score [range: 0 (no deficit); 1 (presence of all deficits)] is calculated, dividing the total sum of points by the number of evaluated domains . Based on previous work demonstrating the good predictive power of 10-TaGA for one-year mortality, we classified participants as having low (0-0.24), medium (0.25–0.49), or high (0.50-1) risk of death .
Functional status. We examined detailed information on functional disability using the Brazilian version of the Older Americans Resources and Services Multidimensional Functional Assessment Questionnaire (BOMFAQ)  and the Katz index . The BOMFAQ (range: 0–30; 30 = worst) evaluates dependency in 15 ADLs and instrumental activities of daily living (IADLs). For each item, a score from zero to two is assigned (0: unable to perform the activity; 1: needs supervision or help to perform the activity; 2: completely independent to perform the activity) .
We additionally used the Katz index to evaluate ADLs (feeding, dressing, bathing, toileting, transferring, and continence). Each activity is scored as either zero (unable to perform the activity) or one (completely independent to perform the activity) [14, 19].
Pain. We screened for pain complaints and their intensity using the 5-point Verbal Descriptor Scale: no pain; mild pain; moderate pain; severe pain; or worst possible pain . Participants identified as having any pain were also asked the following question: "Did the pain occur on most days in the past three months?". A negative answer defined sporadic pain, while a positive answer defined persistent pain [24, 25]. We further investigated participants with persistent pain using the multidimensional Geriatric Pain Measure [26, 27].
Frailty. We defined frailty using the Study of Osteoporotic Fractures (SOF)  index for frailty, and the FRAIL scale [29, 30]. The SOF index includes three items: weight loss of 5% or more; inability to rise from a chair five times; and reduced energy level. The score ranges from zero to three points, and classifies patients as: robust (0 points); pre-frail (1 point); or frail (2–3 points) . The FRAIL scale includes five mnemonic questions on fatigue, resistance, ambulation, illnesses, and loss of weight. The score ranges from zero to five points, and classifies patients as: robust (0 points); pre-frail (1–2 points); or frail (≥ 3 points) [29, 30]. Although the phenotypic criteria for the diagnosis of frailty were not used in our study, recent studies have shown that the SOF index and FRAIL scale have similar performances to predict adverse outcomes in vulnerable older adults [32–34].
Physical performance. We measured gait speed instructing participants to walk 4.5 meters at their usual pace and used the faster of two measurements in our analyses. Participants were allowed the use of assistive devices whenever necessary .
We measured handgrip strength using a Saehan dynamometer. We requested that participants sit on armless chairs, with their spines erect, shoulders positioned in adduction and neutral rotation, elbows flexed at 90°, forearms in half pronation, and neutral wrists. They would then squeeze the device using their dominant hand, applying their maximum strength. We used the mean value of three measurements in our analyses .
Finally, we requested that those who were able to sit and stand independently do the chair-stands test. We asked that participants do their best to complete five sit-to-stand repetitions, without the help of the arms, and recorded the total time in seconds .