Methodology

How the index is built

The construction follows the OECD and EU Joint Research Centre handbook on composite indicators. This page sets out the working design; choices marked open are still being decided and will be versioned when settled.

Draft. No scores have been calculated yet.

01

Framework

The index asks three questions of each country: how people are doing, whether care is available when they need it, and which conditions shape where the country is heading.

Outcomes

Suicide, depression and anxiety, alcohol and drug harm, life evaluation.

Care

Workforce, beds, public spending, treatment coverage, policy and legislation, quality of care.

Conditions

Life satisfaction, social support, unemployment, and the three-year change in each outcome.

Economic indicators are reported alongside the score as context, not inside it. The index aims to show the effect of mental health on economies, so using economic inputs would make that argument circular.

02

Construction

  1. Indicator selection

    Soundness, measurability, coverage and relevance, with proxies documented.

  2. Outliers and imputation

    Gaps are filled only where the method is stated, and every imputed value is flagged.

  3. Multivariate check

    Correlations and principal components tested against the framework.

  4. Normalisation

    Indicators put on a common scale, with direction set so higher always means better.

  5. Weighting and aggregation

    Equal weights within pillars for version 1.0.

  6. Uncertainty and sensitivity

    A Monte Carlo simulation varies imputation, normalisation and weights; each country is published with a rank interval.

  7. Decomposition

    Every score can be broken down by pillar and indicator.

03

Levels and change

A country with a stable high suicide rate is in a different position from one where the rate is rising. The index will report levels and the three-year change side by side.

04

Release cadence

Annual core index

The full score and ranking, recalculated once a year, with a stated rule for carrying forward indicators that have not been updated. Most inputs are annual and arrive two to three years late.

Quarterly pulse

A short release limited to series that move within the year: monthly suicide counts in the Baltic states and Sweden, service data in England and Portugal, and US mortality and overdose deaths.

05

Open decisions

What the index measures

Mental health, mental illness, or the direction a country is heading. Prevalence and service data describe illness; wellbeing and conditions describe health.

Working default: all three pillars, reported separately as well as combined.

Economic indicators

Working default: a separate context layer, outside the score.

Levels, change, or both

Working default: both, side by side.

Publishing entity

Some data licences restrict use connected to a commercial enterprise. Whether the index is published by Paracelsus Recovery or by an independent foundation affects which sources can be used. See licences.

Weighting

Working default: equal weights within pillars for version 1.0, tested by sensitivity analysis. Expert or statistical weights may follow.

Indicator list

What goes in

The full candidate list from the project specification. Status shows whether each publisher's code has been confirmed against the live API. Blocked and permission-needed sources are listed so the gaps are visible.

06

Changelog

  • 0.1-draftFramework, indicator list and open-data pipeline set up; first data from Eurostat and CDC; national sources reviewed for 33 jurisdictions.