Skip to content
Davide Scuteri Moretti
Senior Data Architect
Home/Architectures/Healthcare
Healthcare
Client
Istituti Fisioterapici Ospitalieri (IFO), Rome
Role
Lead Data Architect and Metadata-Driven ETL Designer

Pharmaceutical Logistics Data Programme

A Strategic Information Foundation for IFO

An institutional capability for turning fragmented operational information into a coherent, governed and reusable picture of pharmaceutical and related consumption across IFO.

Architecture Flow
AL Healthcare DataObjects10 logical source domainsMetadata-Driven PythonETL59 field rules · 17transformsFactsconsumi · budget · forecastDimensionsCalendar · Org · Drug · ATC· DeviceAnalytical MartReconciled scenarios
§ 01 Context

IFO is the institutional structure bringing together two IRCCS institutes — the Istituto Nazionale Tumori Regina Elena and the Istituto Dermatologico San Gallicano — and is described in its own training materials as the only Italian hospital organisation including two IRCCS institutes. In a research hospital, logistics is simultaneously a clinical, organisational, economic and scientific problem: a unit of consumption can represent a patient treatment, a diagnostic procedure, a research protocol, a quality-control activity, a device-assisted service or a planned budget commitment. The programme responds to that fragmentation with a controlled analytical domain for actual consumption and related information, populated from governed AL Data Objects. The technical mapping covers 59 field-level rules, 10 logical source domains, 5 target Data Objects and 17 transformation categories.

§ 02 Designed System
Metadata-driven Python ETL
Analytical fact and dimension model
Deterministic business-key hashes
Idempotent MERGE
Explicit treatment of proxy mappings
Source provenance, duplicate control and key resolution as product features
Temporal consistency, reconciliation and exception management
Versioned mappings, classifications and organisational structures
§ 03 Capabilities
  • Actual, budget and forecast scenarios
  • Calendar, organisation, hospital, department, drug, ATC and device dimensions
  • Total Cost = Quantity × Unit Cost, reconciled per scenario
  • Consumption interpreted by time, site, organisational unit, therapeutic and device identity
  • Indicators that can be explained, reconciled and traced rather than merely presented
  • Research-ready governed views prepared without bypassing institutional responsibilities
§ 04 Objects & Stack
Data Objects
  • fact_consumi
  • fact_budget_consumi
  • AL_laboratory_orders
  • AL_laboratory_results
  • AL_lab_panels
  • AL_quality_control_data
  • AL_lab_instruments_devices_EXT
Technologies
PythonAirflowPostgreSQLSpark
§ 05 Outcome

Success is not the production of a technically correct dataset. Success is the routine use of a trusted institutional capability, in which pharmacists, controllers, clinical managers, researchers and academic partners work from the same definitions and understand how results were produced.

§ 06 Institutional Programme
Istituti Fisioterapici Ospitalieri (IFO), Rome — IRE and ISG

To explain the work at institutional and strategic level: why the programme matters to IFO, how it supports hospital pharmaceutical logistics, and why IFO's academic network makes the initiative particularly valuable. Solution design, code and physical schemas remain deliberately outside its scope.

Create a trusted institutional view of consumption

Establish a shared and governed representation of pharmaceutical, device and related consumption across time, sites and organisational units.

Support economic and operational stewardship

Connect quantities and costs with budgets, forecasts and organisational accountability, enabling earlier and better-informed management action.

Strengthen clinical and patient-safety visibility

Improve the ability to investigate patterns, anomalies, unusual consumption and operational dependencies that may affect treatment continuity and safe care.

Increase auditability and reproducibility

Ensure that figures can be traced to authoritative sources, interpreted consistently and regenerated under controlled rules.

Enable research and academic use

Provide a higher-quality foundation for pharmacoeconomic analysis, observational studies, health-services research, technology assessment and student or specialist projects.

Create a reusable institutional capability

Build a model that can grow with IFO, accommodate new therapeutic areas and support future integration without losing semantic control.

§ 07 Institutional Value
Executive and strategic management
  • An integrated view of consumption, cost and budget performance across IFO.
  • Improved visibility of structural trends and exceptional deviations.
  • A stronger basis for resource allocation, investment decisions and service planning.
  • Greater confidence that management indicators can be explained and reproduced.
Hospital pharmacy and pharmaceutical governance
  • Consistent visibility of products, therapeutic classes, quantities, units and costs.
  • Improved support for stock, consumption and expenditure analysis.
  • Better identification of unusual patterns, substitution effects and data-quality issues.
  • A basis for more structured dialogue with clinical departments and procurement.
Clinical departments and care pathways
  • Clearer linkage between consumption and organisational or clinical context.
  • Improved capacity to compare periods, sites and pathways without relying on local spreadsheets.
  • Support for multidisciplinary review of therapies, devices and associated operational impacts.
  • Greater transparency around the resource implications of complex care.
Finance, management control and procurement
  • Coherent reconciliation of actuals, budgets and forecasts.
  • Improved analysis of price, quantity and mix effects.
  • More reliable preparation of procurement scenarios and contract discussions.
  • Faster investigation of discrepancies and clearer audit trails.
Research leadership and investigators
  • Reusable, documented and reproducible data assets for health-services and pharmacoeconomic studies.
  • Improved readiness for multicentre research and joint projects with universities.
  • Better support for cohort construction, protocol feasibility and operational research.
  • A stronger basis for demonstrating the real-world context in which therapies are delivered.
Students, residents and academic partners
  • Access to a mature institutional case study in healthcare data governance.
  • Opportunities for theses, internships, specialist training and joint methodological work.
  • A practical environment in which clinical, economic and data-science perspectives can meet.
  • Exposure to the governance standards required in an IRCCS setting.
§ 08 Governance & Use
Governance principles
  1. 01Clinical meaning before technical convenience.
  2. 02Definitions that reflect pharmaceuticals, devices, organisational units, time periods and scenarios as they really are.
  3. 03Traceability of every published figure to authoritative sources.
  4. 04Data quality treated as part of the product, not as an afterthought.
  5. 05Versioned change: mappings, classifications and organisational structures documented so historical analyses remain intelligible.
  6. 06Academic usability without loss of control: research-ready governed views rather than bypassed responsibilities.
  7. 07Sustainable ownership across pharmacy, management, clinical governance, research and data architecture.
Representative use cases
Budget monitoring and early variance analysis

Deviations identified before period close rather than reconstructed afterwards.

Executive reporting

Consistent indicators that can be explained, reconciled and traced rather than merely presented.

Benchmarking and network collaboration

Comparable definitions and evidence prepared for regional, national or multicentre initiatives.

Academic and translational research

Governed views supporting pharmacoeconomics, health-technology assessment, pharmacoepidemiology and pharmacovigilance with university partners.

What success should look like
AreaIndicatorDirection
ResearchFeasibility studies, theses or joint analyses supportedIncrease
AuditabilityShare of indicators with documented provenance and rule versionIncrease
PlanningBudget deviations identified before period closeIncrease

In a modern research hospital, excellence depends not only on possessing data, but on being able to explain what those data mean, where they came from, how they can be trusted and how they should be used.

§ 06 Related Reading
Essay 24
Data Quality Is a Computational Property
*Data quality is often organised as a terminal inspection: after ingestion and transformation, a suite of checks decides whether the dataset is good. This essay argues that quality is not an external label attached to a finished product but a computational property produced, transformed, and sometimes degraded at every stage. It develops a multidimensional quality model, distinguishes intrinsic, contextual, representational, and process quality, and formalises quality propagation through tasks. It addresses uncertainty, thresholds, fitness for purpose, and the dangers of composite scores. The philosophical claim is that “correct data” has no universal meaning independent of use. Quality must be specified relative to the claims a dataset is expected to support, while still preserving objective constraints such as identity, units, and temporal coherence.*
Essay 17
The Task as the Atomic Unit of Data Computation
*The task is often treated as an operational convenience: a box in an orchestrator, a scheduled job, a notebook, or a container invocation. This essay argues that the task should instead be understood as the atomic unit of accountable data computation. A task is not merely a step that runs; it is a typed transformation with declared inputs, outputs, preconditions, postconditions, quality obligations, provenance effects, and failure semantics. The essay formalises tasks as state transitions and morphisms over data contracts, distinguishes logical tasks from physical executions, examines composability in directed acyclic graphs, and relates task design to ideas from programming language semantics and the philosophy of action. The central claim is that a platform becomes intelligible when its smallest executable unit is also its smallest explainable unit. Granularity is therefore not only a performance concern. It determines whether lineage, replay, ownership, and change impact can be reasoned about without reconstructing intent from code after the fact.*