Implementing Minimum Viable Data Sets (MVDS) in Healthcare Improvement

Why this topic matters

Data collection is fundamental to clinical audit, quality improvement (QI), and service evaluation across the NHS. Without relevant and reliable data, it’s impossible to understand current performance, identify areas for improvement, or measure the impact of interventions. However, the inclination to collect 'everything just in case' often leads to several problems:

• Increased Clinician Burden: Extensive data fields take valuable time away from direct patient care or other clinical duties.
• Poor Data Quality: When data collection is burdensome, accuracy and completeness can suffer. Unclear definitions or too many optional fields lead to inconsistent recording.
• Analysis Paralysis: Large datasets, especially those with many irrelevant fields, can be difficult and time-consuming to analyse, delaying insight and action.
• Resource Inefficiency: Storing and managing large volumes of data requires significant IT infrastructure and staff time, diverting resources from other areas.
• Delayed Improvement Cycles: The time spent on excessive data collection and analysis prolongs improvement cycles, slowing the pace at which positive changes can be implemented.

Implementing an MVDS approach helps to mitigate these issues, making data collection more purposeful, sustainable, and impactful for improvement initiatives.

Practical explanation: What is a Minimum Viable Data Set?

A Minimum Viable Data Set (MVDS) is the smallest collection of data points necessary to achieve a specific objective. This objective could be to:

• Answer a specific clinical audit question.
• Monitor the key outcome of a quality improvement project.
• Evaluate the effectiveness of a new service pathway.
• Track compliance with a national guideline.

The 'minimum viable' aspect is crucial. It’s about stripping away non-essential elements to focus on the signal rather than the noise. The concept draws parallels with the 'Minimum Viable Product' in software development, where the aim is to launch a product with just enough features to satisfy early customers and provide feedback for future development.

Key characteristics of an effective MVDS:

• Purpose-driven: Each data element is directly linked to a specific question or objective.
• Concise: Contains the fewest possible data points.
• Clear Definitions: Every data point has an unambiguous definition to ensure consistent interpretation and collection.
• Feasible: Data can be reliably collected with existing resources and systems, minimising additional burden.
• Actionable: The data collected provides insights that can directly inform decisions and actions for improvement.
• Evolvable: An MVDS is not static; it can be reviewed and revised as objectives change or more comprehensive data becomes necessary.

Common pitfalls in data collection without an MVDS approach

Without a structured approach like MVDS, teams often encounter several common challenges:

• 'Just in case' collection: Gathering data without a clear purpose, in the hope it might be useful later.
• Legacy data fields: Continuing to collect data points that were once relevant but are no longer used or needed.
• Lack of standardisation: Different individuals or teams collecting the same data point in varying ways, making aggregation and comparison difficult.
• Over-reliance on free-text fields: While sometimes necessary, excessive free text can be hard to analyse systematically.
• Insufficient stakeholder engagement: Failing to involve those who will collect or use the data in the design phase, leading to resistance or poor adoption.
• Ignoring data validity and reliability: Collecting data without considering if it accurately reflects the intended measure or if it can be consistently collected over time.

Step-by-step approach to defining and implementing an MVDS

Implementing an MVDS requires a structured, collaborative process. Here’s a practical framework:

Step 1: Define your objective(s)

Clearly articulate what you want to achieve or answer. What specific problem are you trying to solve? What hypothesis are you testing? What service are you evaluating? Use frameworks like SMART (Specific, Measurable, Achievable, Relevant, Time-bound) to refine your objectives.

• Example: "Reduce the average time from Emergency Department (ED) arrival to consultant review for patients with suspected sepsis by 30 minutes within six months." or "Assess compliance with national guidelines for venous thromboembolism (VTE) prophylaxis in elective orthopaedic patients."

Step 2: Brainstorm potential data points

Bring together a multidisciplinary team, including clinicians, data collectors, QI leads, and potentially a data analyst. Brainstorm all possible data points that could be relevant to your objective(s). Don't filter or judge at this stage.

Step 3: Prioritise and narrow down to the 'minimum viable'

Review your brainstormed list against your objective(s). For each data point, ask:

• Is this absolutely essential to answer my objective? (If not, discard it).
• Can this data point be reliably and practically collected? (Consider existing systems, resources, and burden).
• Is there an existing, reliable source for this data? (Leverage electronic patient records, national databases where possible).
• Is this data actionable? Will knowing this information lead to insights that inform improvement?

This filtering process is the core of creating an MVDS. Be ruthless in eliminating anything that isn't strictly necessary.

Step 4: Define each data element unambiguously

For each selected data point, create clear, concise definitions. This includes:

• Data element name: e.g., 'Time to Consultant Review'
• Definition: 'The time elapsed from patient registration in ED until the initial face-to-face assessment by an ED Consultant or registrar with consultant sign-off.'
• Format: e.g., 'HH:MM' or 'Date (DD/MM/YYYY)'
• Units (if applicable): e.g., 'minutes', 'mg/kg'
• Allowed values/range: e.g., 'Yes/No', '0-100', 'referral source dropdown'
• Source: Where will this data be extracted from?

This step ensures consistency, reducing variations in how data is interpreted and recorded.

Step 5: Design the collection method and tool

Based on your MVDS, design a pragmatic data collection method. This could be:

• A focused section within an existing electronic patient record (EPR).
• A digital form (e.g., using a secure NHS-approved platform).
• A bespoke spreadsheet with clear validation rules.
• Leveraging existing coded data for extraction.

Minimise manual transcription and aim for direct entry where possible. Pilot testing is crucial here.

Step 6: Pilot, review, and iterate

Before full-scale implementation, pilot your MVDS and collection method with a small sample. Gather feedback from those collecting the data and those using it.

• Is the data collection feasible and clear?
• Are there any ambiguities in the definitions?
• Does the collected data actually answer your objectives effectively?

Be prepared to refine your MVDS based on this feedback. An MVDS is a living document, not a fixed entity.

Step 7: Train and communicate

Ensure all relevant staff are trained on the MVDS definitions and collection process. Clearly communicate the why behind the MVDS – how it will reduce burden and drive meaningful improvement. Provide ongoing support and opportunities for feedback.

Example in clinical practice: Hip Fracture Pathway QI Project

Old approach (pre-MVDS):

A hospital aimed to improve its hip fracture pathway. They collected data on over 50 different points for every patient, including detailed anaesthetic notes, full pre-op assessments, all allied health professional interventions, and extensive follow-up questionnaires. This led to:

• Junior doctors spending significant time filling out a paper proforma.
• Data quality issues due to fatigue and unclear fields.
• A large dataset that was difficult to analyse, delaying progress reports.

MVDS approach for specific QI objective: Reduce 'Time to Theatre' for hip fracture patients to less than 36 hours

Objective: To improve adherence to the 36-hour 'Time to Theatre' standard for hip fracture patients, as recommended by NICE guidelines and national audits (e.g., NHFD - National Hip Fracture Database).

Key questions:

1. What is the current 'Time to Theatre' performance?
2. What are the key delays contributing to non-compliance?
3. Is our intervention (e.g., new theatre scheduling, enhanced pre-assessment) improving this metric?

MVDS defined (after multidisciplinary agreement):

• Patient Identifier (Local Hospital Number): For linking data.
• Date/Time of Arrival in ED: When patient arrived in hospital (from EPR).
• Date/Time of Diagnosis of Hip Fracture: Confirmed by imaging (from EPR radiology report).
• Date/Time 'Fit for Surgery' Confirmed: Clinical decision documented (from EPR anaesthetic/surgical notes).
• Date/Time of Incision (Theatre Start Time): When surgery commenced (from theatre system).
• Ward for Post-Op Care: To track pathway adherence.
• Reason for Delay (>36 hours): Single choice from a dropdown list (e.g., 'Patient comorbidities/optimisation needed', 'Anaesthetic availability', 'Theatre capacity', 'Senior clinical decision'). This is crucial for identifying actionable barriers.
• 30-Day Mortality: For safety monitoring (extracted from central system).

This MVDS dramatically reduced the data collection burden, improved the completeness and accuracy of the critical data points, and allowed the QI team to rapidly identify the primary causes of delay. Weekly reports could be generated swiftly, enabling focused interventions and demonstrating tangible improvements in patient care within months.

How Lazomis can help

Lazomis provides a robust and secure platform that can significantly aid in the implementation and management of Minimum Viable Data Sets for your improvement projects, clinical audits, and service evaluations.

• Customisable Data Collection Forms: Easily build digital forms tailored to your MVDS, ensuring only the essential data points are collected. Our drag-and-drop interface helps you define fields with clear instructions, dropdowns, and validation rules, embodying your MVDS definitions.
• Secure Data Storage and Access: All data is held securely, adhering to NHS data governance standards, making it safe for sensitive clinical information.
• Automated Data Visualisation: Once your MVDS data is collected, Lazomis dashboards can automatically generate real-time visualisations, helping you track your key metrics without delay. This supports rapid feedback and iteration cycles.
• Team Collaboration: Facilitate multidisciplinary team involvement by allowing different roles to contribute to data collection or review, fostering a shared understanding of the MVDS and its purpose.
• Audit Trail and Version Control: Maintain transparent records of data collection activities and revisions to your MVDS, promoting good governance.
• Integration Potential: While always recommending local IT/informatics guidance, Lazomis is designed with potential for integration with other systems to streamline data flow where feasible, reducing manual entry.

This resource supports, but does not replace, clinical judgement. Local policy, formulary and specialist advice should be followed.