Team Poop Poop
Inspiration
Singapore is ageing fast, and dementia is one of the biggest drivers of long-term community care demand. The 2024 Lancet Commission on dementia identifies hearing loss as one of the two largest modifiable risk factors for dementia, linked to about 7% of cases. Untreated hearing loss raises cognitive load and drives social isolation - both linked to cognitive decline.
Three major studies shaped our idea:
- Slower cognitive decline - ACHIEVE trial (The Lancet, 2023). In this randomised controlled trial of adults aged 70-84 with hearing loss, hearing intervention showed no significant effect across the whole group, but among higher-risk older adults it was associated with ~48% slower cognitive decline over 3 years.
- Lower risk of decline and dementia - Yeo et al., NUS Yong Loo Lin School of Medicine (JAMA Neurology, 2022). A meta-analysis of 31 studies (137,484 participants) found hearing-aid users had 19% lower risk of long-term cognitive decline, and hearing aid use was associated with lower risk of developing dementia.
- Earlier is better - Framingham Heart Study (JAMA Neurology, 2025). Over up to 20 years of follow-up, adults with hearing loss who started hearing aids at age 60-69 had 61% lower dementia risk; no significant association was seen for those starting at 70+.
Together, these suggest that hearing support - especially when started early and targeted at higher-risk seniors - can slow cognitive decline, is associated with lower dementia risk, and may delay dementia onset.
The gap in Singapore today
- Hearing loss is common: in the Project Silver Screen pilot, 7 in 10 seniors screened had some form of hearing loss (MOH, 2018).
- Screening already exists: MOH's Project Silver Screen (run by HPB since 2018) offers hearing, vision and oral screening for seniors 60+ at S$0-S$5, and refers those with hearing problems for further assessment.
- Funding already exists: AIC's Seniors' Mobility and Enabling Fund (SMF) covers up to 90% of hearing-aid costs for citizens and PRs aged 60+ (per capita household income up to S$4,800). Yet only ~4,000 seniors a year have applied for hearing-aid subsidies since 2019 (MOH, 2024).
- Philanthropic top-ups have ended: Temasek Foundation Cares managed S$35M from Temasek and 11 corporate donors to fund Project Silver Screen vouchers (S$15-S$200) for hearing aids and other aids - but these vouchers ceased on 1 Oct 2024.
- The gap is follow-through, not screening: despite 7 in 10 screened seniors having some hearing loss and subsidies of up to 90%, relatively few go on to get hearing aids.
Our question: if follow-through from screening to hearing aids were improved and better targeted - by age, risk and planning area - how much future dementia care demand could Singapore avoid, and is it worth the money?
Why does even a one-year delay matter? Each extra year a senior stays independent means:
- Lower government healthcare spending on community care, nursing homes and subsidies
- Lower family costs: fewer caregiver hours, less spending on domestic helpers and care services
- Less GDP loss: fewer family members cutting work hours or leaving jobs to become caregivers
What it does
Problem statement: A3 - HealthPulse: Community healthcare demand forecasting
HearWell MindWell is a proposed community care forecasting platform with a built-in hearing-health policy simulator:
- 5-year demand forecast - projects community care demand (eldercare and dementia-related) for every planning area.
- Supply vs demand gap analysis - compares projected demand against existing facilities and beds in an interactive dashboard.
- Top 5 underserved areas - a map flagging where infrastructure investment is needed most.
- Policy simulator - models the pathway: hearing screening → hearing-aid uptake → slower cognitive decline → lower dementia risk / delayed onset → avoided care demand → savings for government, families and the economy. Includes optimistic / base / pessimistic toggles based on the effect sizes above.
- Budget optimiser - under a fixed budget, recommends which age groups (e.g. 60-69, per Framingham), risk groups and planning areas to prioritise.
- Policy scenarios - tests options such as:
- Improving follow-through from Project Silver Screen to hearing assessment and SMF-subsidised hearing aids
- Prioritising outreach for seniors aged 60-69, where evidence suggests earlier hearing-aid use matters most
- Linking Project Silver Screen results to seniors' Healthier SG family doctors, so hearing follow-up becomes part of their health plan
- Reinstating top-up vouchers through philanthropic partners such as Temasek Foundation
The key output answers: for every S$1 spent on targeted hearing support, how much could slowing decline and delaying dementia save, and when does the programme break even? Even if universal coverage is not cost-effective, the model shows whether early, targeted intervention is.
How we plan to build it
- Ingestion: open datasets from data.gov.sg, SingStat, MOH and HDB loaded into Databricks using a Bronze → Silver → Gold medallion pipeline (raw data → cleaned and standardised data → analysis-ready tables).
- Governance: Unity Catalog for data lineage, access controls, and clear tagging of source data vs modelled assumptions.
- Forecasting: time-series and regression models in Databricks ML, tracked with MLflow, projecting demand by planning area from population ageing and admissions trends.
- Simulation and optimisation: a cost-effectiveness model valuing each year of delayed dementia onset (care costs, caregiver costs, productivity loss), using published SMF subsidy rules, plus an operations research (linear programming) layer that allocates a fixed budget across age groups, risk groups and areas.
- Dashboards: Databricks SQL dashboards, with Genie so policymakers can ask questions in plain English (e.g. "Which 5 areas gain most from a S$1M hearing programme for 60-69-year-olds?").
Challenges we foresee
- Different strengths of evidence: slower cognitive decline comes from a randomised trial (ACHIEVE, higher-risk subgroup); lower dementia risk comes from observational studies (association, not proven causation); delayed onset is inferred from both. We will model effect sizes as scenarios and will not claim hearing aids prevent dementia or reverse it in people who already have it.
- Data gaps: there is no open dataset on hearing-loss prevalence or screening follow-up by area, and HDB elderly data is published by town rather than planning area, so we will estimate area-level figures from national figures and age structure, and state these assumptions clearly.
- Valuing a year of delay: combining government care costs, family caregiving costs and productivity loss into one consistent measure across sources with different years and granularities.
What we learned during the ideation stage
- From reviewing the research, targeting and timing appear to matter more than coverage: the reported benefits are concentrated in higher-risk seniors and in those who start hearing aids before 70.
- From reviewing Singapore's current policies, screening (Project Silver Screen) and funding (AIC SMF, up to 90%) are already in place. The gap appears to be follow-through - many seniors screened have hearing loss, but relatively few go on to obtain subsidised hearing aids (~4,000 a year).
- The strength of evidence varies across studies (a randomised trial vs observational studies), which shaped our decision to model outcomes as scenarios rather than fixed predictions.
- Even a one-year delay in dementia onset could have value beyond healthcare, for families and for the workforce, which is why our model will measure caregiver costs and productivity loss, not just government spending.
Possible modifications to the current proposed model
- Add social-isolation and Active Ageing Centre data to refine risk scoring.
- Partner with AIC, HPB, polyclinics and philanthropic partners such as Temasek Foundation to validate assumptions with real programme data.
- Extend the simulator to other modifiable dementia risk factors in the Lancet Commission, such as vision loss and physical inactivity.
References
- Lin FR, et al. Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE): a multicentre, randomised controlled trial. The Lancet, 2023.
- Yeo BSY, et al. Association of Hearing Aids and Cochlear Implants With Cognitive Decline and Dementia: A Systematic Review and Meta-analysis. JAMA Neurology, 2022 (online Dec 2022; print 2023). NUS Yong Loo Lin School of Medicine.
- Framingham Heart Study analysis of hearing aid use and incident dementia. JAMA Neurology, 2025.
- Livingston G, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet, 2024.
Singapore policy sources
- MOH. More Seniors to Benefit from Functional Screening (Project Silver Screen), 19 Sep 2018.
- MOH Parliamentary Reply. Statistics of successful applications by seniors to AIC for funding of hearing aids, 11 Nov 2024.
- MOH Parliamentary Reply. Hearing loss data in NPHS and auditory tests under Healthier SG, 12 Jan 2026.
- AIC. Seniors' Mobility and Enabling Fund (SMF) eligibility.
- Temasek Foundation. Project Silver Screen.
Built With
- auto-loader
- dashboards
- data-gov-sg
- databricks
- databricks-sql
- delta-lake
- genie
- geospatial
- healthcare
- jupyter-notebook
- linear-programming
- machine-learning
- mlflow
- operations-research
- pyspark
- python
- sql
- time-series-forecasting
- unity-catalog
Log in or sign up for Devpost to join the conversation.