Consultations & Clinical Documentation

The consultation is where medicine actually happens. Everything else - scheduling, billing, letters, recalls - exists to support the clinical encounter at the centre. How your software handles consultations determines the quality of your clinical record, the efficiency of your clinicians, and your ability to audit, report on, and learn from the care you deliver.

This comparison examines how Jump EHR and Semble approach the clinical consultation: how notes are structured, what data is captured, how templates work, what AI assistance is available, and what the resulting clinical record looks like. The differences here are not incremental. They represent fundamentally different philosophies about what a consultation record should be - and those philosophies have consequences for clinical governance, data quality, and day-to-day efficiency that compound over every patient encounter.

At a Glance: The Headline Differences

• Clinical record model - Jump: Problem-oriented (POMR) with each problem documented separately. Semble: Section-based free text with optional templates.
• SNOMED CT coding - Jump: Every problem, finding, observation, procedure, and allergy is SNOMED-coded. Semble: Not documented - no SNOMED coding in consultations.
• Structured observations / vitals - Jump: SNOMED-coded with UCUM units, reference ranges, abnormality flags, derived scoring (e.g. PHQ-9). Semble: Not documented as structured data capture.
• Problem episode tracking - Jump: New Episode, Review, Flare Up, End Episode per problem. Semble: Active / Inactive status only.
• Allergy recording in consultation - Jump: Structured with SNOMED allergen, criticality, SNOMED-coded reaction manifestations, culprit substance. Semble: Free text allergy section (shows in summary by default).
• Pre-built consultation templates - Jump: System templates for common clinical scenarios (guideline-based). Semble: No pre-built templates - create from scratch.
• Template item types - Jump: 6 structured types including findings, observations, social history, procedures, medication review, immunisation review. Semble: Free text sections with custom questions (required/optional).
• Template conditional logic - Jump: Show/hide items based on captured values (equals, exists, greater than, etc.). Semble: Not documented in consultation templates.
• Template reusable blocks - Jump: Packaged item sets shareable across templates; update once, propagate everywhere. Semble: Not available.
• Template immutability - Jump: Full snapshot preserved per consultation - immune to later template edits. Semble: Not documented.
• Clinical intent tagging - Jump: Screening, monitoring, diagnostic, safety net, administrative per item. Semble: Not available.
• Item priority levels - Jump: Core / Recommended / Optional per template item. Semble: Required / Optional questions only.
• AI consultation tool - Jump: Built-in freeform text to SNOMED-coded problems, observations, allergies, tasks, document suggestions. Semble: Heidi AI ambient scribing (separate product at extra cost).
• AI reads clinical record - Jump: Yes, deduplicates against existing problems and allergies. Semble: No, Heidi reads consultation transcript only.
• Voice transcription - Jump: Built-in with UK medical terminology vocabulary, included. Semble: OS-level dictation or Heidi AI (extra cost).
• Consultation medium tracking - Jump: Face-to-face, telephone, video, email, SMS - each SNOMED-coded. Semble: Consultation type selectable.
• Clinical setting - Jump: GP practice, home visit, care home, community clinic, remote - SNOMED-coded. Semble: Not documented as structured field.
• Procedures in consultation - Jump: SNOMED-coded with status, body site, not-done reasons. Semble: Not documented as structured.
• Family history capture - Jump: Structured panels with SNOMED-coded conditions per family relationship. Semble: Not documented as structured.
• Immunisation review - Jump: Structured with expected status, prior history, reviewed outcome. Semble: Not documented in consultation workflow.
• Medication review prompts - Jump: Structured with intent (review/continue/consider/stop), clinician response. Semble: Not documented as structured prompts.
• Referral capture in consultation - Jump: SNOMED-coded services with FHIR priority (routine/urgent/asap/stat), performer, organisation. Semble: Not documented in consultation - via Letters module.
• Consultation audit trail - Jump: Full amendment history with required reason, preserving original record. Semble: Edit history tracks clinician name change.
• Auto-save drafts - Jump: Continuous auto-save throughout consultation. Semble: Auto-save (same device/browser only).
• Write notes on behalf of clinician - Both platforms support this.
• Consultation diagrams - Jump: Not documented. Semble: Available for annotating anatomical diagrams.
• Video consultations - Jump: Supported (remote delivery mode). Semble: Built-in with virtual waiting room, background blur, noise cancellation, up to 200 participants.
• 75+ specialty workflows - Jump: Configurable via templates. Semble: Marketed as supporting 75+ specialties.

The Clinical Record Model: Problem-Oriented vs Section-Based

This is the most consequential architectural difference between Jump and Semble's consultation modules. It shapes everything else - how data is captured, how it's coded, how it's queried, and how it's used for clinical governance.

Jump: Problem-Oriented Medical Records (POMR)

Jump's consultation mode is built around problem-oriented medical records. Every consultation is structured by the clinical problems being addressed, not by a single flat text entry. Each problem gets its own set of clinical documentation sections (history, examination, assessment, plan, and others), its own structured data entries, and its own SNOMED CT code. A consultation addressing three problems - hypertension review, new presentation of knee pain, and medication side effect - produces three distinct, coded, queryable clinical records.

Each problem is assigned an episode type: New Episode, Review, Flare Up, or End Episode. This gives the practice a structured longitudinal view of how conditions progress over time. A "Review" of asthma is clinically different from a "Flare Up" of asthma, and the record captures that distinction.

Problems can be drawn from the patient's existing problem list (continuing an existing clinical thread) or created new during the consultation. The SNOMED CT code is assigned at the problem level, meaning every diagnosis, symptom, and clinical concern is terminologically standardised.

Semble: Section-Based Free Text

Semble's consultation notes use a section-based approach. Clinicians create a consultation note for a patient, select a consultation type (e.g., "doctor consultation"), optionally choose a template, and then write into sections. The default sections include Problem Heading, Medical History, Allergies, and others that can be added via tags at the bottom of the note. Sections can be toggled to "Show in Summary" to populate the patient's overview page.

The Problem Heading and Medical History sections support Active/Inactive status assignment, helping track which conditions are current. Allergies appear in the patient's summary by default for safety. Semble auto-saves drafts during writing (on the same device and browser only).

This approach is simpler and more familiar - it resembles traditional clinical note-taking. But it means the consultation record is fundamentally a text document with optional structure, rather than a structured clinical dataset with narrative context.

Why the Difference Matters

The practical consequences of this architectural choice cascade through every aspect of clinical record-keeping:

Querying and reporting. With POMR, a practice can ask "show me all patients with a diabetes review in Q4 where HbA1c was above 64 mmol/mol." With section-based free text, this requires manually reading every consultation note. The difference is between a database query and a document search.

Clinical audit. CQC inspections and clinical governance reviews increasingly expect practices to demonstrate evidence-based care through auditable data. SNOMED-coded problems, observations, and procedures are auditable by definition. Free text notes require manual review.

Longitudinal patient records. When every consultation contributes SNOMED-coded problems with episode types, the patient's record builds into a structured clinical timeline - not just a chronological list of notes, but a queryable history of coded diagnoses, observations, and outcomes.

Interoperability. SNOMED CT is the NHS standard for clinical terminology. SNOMED-coded records can be shared with other systems (NHS GP systems, hospitals, referral pathways) in a format that preserves clinical meaning. Free text records can be shared as documents but lose their clinical structure in transit.

Structured Data Capture: What Gets Coded

Jump's consultation captures structured data across ten clinical header types within each problem. Every piece of structured data is linked back to both the consultation and the specific problem, enabling granular clinical queries.

Observations and Vitals

Jump captures observations as SNOMED-coded clinical measurements with UCUM (Unified Code for Units of Measure) units and reference ranges. Blood pressure, heart rate, respiratory rate, oxygen saturation, temperature, weight, height, BMI, peak flow, HbA1c, eGFR - whatever the clinical scenario requires. Each observation carries abnormality flags and can include a comment. Derived observations support auto-computation for questionnaire scoring: a PHQ-9 total score auto-computes from the nine individual question values, updating in real time as source values are entered. Observations recorded in consultations become part of the patient's longitudinal observation record, queryable and trendable over time.

Semble's consultation notes do not document a structured vitals or observations capture system within the consultation workflow. Vital signs would be recorded as free text within note sections or captured through Semble's separate patient record fields. The distinction is between observations that are coded, measured, trended, and queryable versus observations written into a narrative note.

Allergies

Jump records allergies as fully structured clinical data: SNOMED-coded allergen, clinical status, verification status (unconfirmed, provisional, confirmed, refuted, or entered-in-error), type, category, criticality (high/low/unable to assess), onset date, and detailed reaction records. Reactions are individually SNOMED-coded from a curated library of manifestation types covering anaphylaxis, angioedema, respiratory reactions, dermal reactions, and gastrointestinal reactions. Allergies recorded during a consultation appear immediately in the patient's allergy list and are available across the clinical record. The five-level verification status is a clinical safety feature - distinguishing between a confirmed penicillin anaphylaxis and an unconfirmed patient-reported intolerance changes prescribing decisions.

Semble captures allergies in a dedicated consultation section that appears in the patient's summary by default for safety. However, the level of allergy structuring (coded allergens, reaction types, criticality classification) is not documented in Semble's public help centre. This is an important clinical safety distinction - structured allergy data enables automated safety checks; narrative allergy notes rely on the prescribing clinician reading and interpreting text.

Procedures, Family History, and Social History

Jump captures procedures as SNOMED-coded events with full status tracking (eight statuses from preparation through completion, including on-hold, stopped, and entered-in-error), body site and laterality recording, and not-done reasons. Family history is recorded in structured panels with SNOMED-coded conditions linked to specific family relationships. Social history covers smoking, alcohol, employment, living situation, and exercise with coded answers, free text, and quantities as appropriate.

Semble's consultation sections can include free text content for any of these clinical areas, but structured, coded capture for procedures, family history, and social history within the consultation workflow is not documented in Semble's public help centre.

Consultation Templates: Structured Guides vs Question Sets

Jump's Template System

Jump's consultation templates are SNOMED-coded, clinically structured guides that follow clinical guidelines (NICE, GOLD, KDIGO) and define the observations, findings, procedures, and review prompts relevant to a specific clinical scenario. When a template is applied, it presents the clinician with a structured data entry form tailored to that consultation type. This isn't a text scaffold - it's a live data capture tool that feeds directly into the patient's coded clinical record.

Templates support six structured item types: clinical findings (SNOMED-coded observations with value types), observations (vital signs with UCUM units and reference ranges), social history (structured lifestyle data), procedures (SNOMED-coded with status tracking), medication review prompts (structured intent and response), and immunisation review prompts (structured vaccine review with prior history).

Template intelligence goes further. Conditional visibility shows or hides items based on captured values - if smoking status is "current," the pack-year history field appears; if a screening score crosses a threshold, additional assessment items become visible. Clinical intent tagging marks every item as screening, monitoring, diagnostic, safety net, or administrative. Priority levels (core, recommended, optional) help clinicians focus when time is short. Reusable blocks - packaged item sets for vital signs panels, cardiovascular examinations, or mental health screening batteries - can be shared across templates and updated centrally.

Jump ships with pre-built system templates for common clinical scenarios. Practices also create custom templates, with version tracking and SNOMED CT document type classification.

Template immutability is a medico-legal safeguard: when a consultation is created, the entire template state is snapshotted - every item, section, and condition. This snapshot is permanently attached to the consultation record. If the template is later edited, historical consultations are unaffected. The record reflects exactly what the clinician was asked to capture at the time.

Semble's Template System

Semble's consultation templates are question-based forms created in Settings. Practices build templates with sections and questions, marking questions as required or optional. Templates are selected when starting a consultation note. The system supports adding sections via tags during note-taking (Problem Heading, Medical History, Allergies, and others), and the "Show in Summary" toggle determines what appears on the patient overview.

Semble's approach is more accessible - creating a template is straightforward, and the freeform nature means clinicians aren't constrained by predefined structures. For specialties where note-taking is highly narrative (psychiatry, psychotherapy), this flexibility has genuine value.

However, Semble's templates do not capture structured clinical data in the way Jump's do. Template responses are essentially answers to questions stored as text, not SNOMED-coded clinical findings with value types, reference ranges, and conditional logic. There are no reusable blocks, no conditional visibility based on captured values, no clinical intent tagging, and no template immutability for medico-legal protection.

AI Consultation Tools: Record-Aware vs Transcript-Based

Jump's Built-In AI

Jump includes an AI consultation tool that transforms freeform clinical text - typed or voice-transcribed - into fully structured, SNOMED-coded clinical data. The clinician writes or dictates their notes naturally, and the AI extracts: multiple distinct clinical problems (up to five per consultation) each with SNOMED code candidates, observations and vitals with codes and units, structured allergy records with coded reactions and criticality, actionable tasks (referrals, investigations, follow-ups, prescriptions) with urgency and due dates, and document suggestions when a referral letter or discharge summary is implied.

The AI processes text against the patient's existing clinical record - active problems and allergies are fetched from the database so the AI can deduplicate intelligently, avoiding duplicate entries. If the patient already has "Asthma" on their problem list and the clinician dictates an asthma review, the AI recognises this as a review of an existing problem, not a new diagnosis.

The output is presented in a structured preview where the clinician includes or excludes individual problems, selects preferred SNOMED codes from multiple AI-suggested candidates (validated against the NHS Terminology Server), edits any section, and reviews extracted data before committing. Nothing is saved without clinician approval. Every AI interaction is audit-logged, and usage is tracked per organisation with a default allowance of 500 AI credits per month (automatically resetting monthly), providing cost visibility and governance.

Clinical safety is built into the AI's design: red flags are always preserved and surfaced prominently, uncertainty qualifiers ("possible," "rule out," "?") are maintained exactly as expressed, no content is invented beyond what the clinician stated, and SNOMED codes favour specificity (avoiding overly generic codes).

Semble's Approach: Heidi AI

Semble's AI capability comes through its Heidi AI integration - a separate ambient scribing product at extra cost per licence per year (after a 30-day free trial). Heidi listens during the consultation, transcribes the conversation, and generates structured notes from 25+ template types. Content can be pushed into Semble's consultation notes or letters.

Heidi is a capable product - clinicians report meaningful time savings, and the integration with Semble means generated notes appear in the patient record. Semble's product page describes the integration as allowing clinicians to "focus on the consultation" while notes are handled in the background.

The fundamental difference is what the AI has access to:

Jump's AI reads the structured clinical record. It knows the patient's complete problem list, medication history, allergies, and previous observations. This enables deduplication, contextual coding, and clinically informed structuring.

Heidi reads the consultation transcript. It knows what was discussed in this session but does not have access to the patient's full record. It cannot deduplicate against existing problems, check historical medications, or contextualise findings against longitudinal data.

Jump's AI produces SNOMED-coded structured data. Problems, observations, allergies, and procedures are output as coded clinical records ready to commit to the structured database.

Heidi produces narrative clinical notes. The output is well-structured text (SOAP notes, specialty templates, letter drafts) but it is text, not coded data. It does not feed into a structured, queryable clinical record in the same way.

Cost: Jump's AI is included in the platform. Heidi costs extra per licence per year, on top of the Semble subscription.

Voice Transcription

Jump includes native voice-to-text transcription optimised for UK clinical language, using speech recognition with a custom medical terminology vocabulary covering NHS clinical terms, standard abbreviations (PMH, DH, SH, FH, O/E), common conditions, medications, examination findings, and red flag symptoms. Recording runs directly in the browser with echo cancellation and noise suppression. Transcribed text feeds directly into the AI consultation tool, creating a seamless dictate-to-structured-record workflow.

Semble's options are OS-level browser dictation (no medical vocabulary), Heidi AI at additional cost per licence, or theoretically Dragon Medical (not officially supported). Jump's transcription is integrated, medically optimised, and included in the platform.

The Consultation Workflow: Step by Step

Jump's Guided Workflow

Jump's consultation follows a guided process: patient selection (or pre-selection from the schedule), consultation setup (medium, clinical setting, date/time, clinician, location - each SNOMED-coded), problem documentation (add problems from the existing list or create new, document clinical sections, apply templates, or use the AI tool), review (all documented problems, linked medications, allergies, observations, procedures, referrals, and documents), and save with full audit trail. Consultations auto-save as drafts throughout.

Each problem within the consultation receives its own set of clinical sections: History, Examination, Observations, Allergies, Social History, Family History, Procedures, Immunisation, Assessment, Plan, Comment, Document, and Follow-up/Review. This means a multi-problem consultation - the norm in general practice - produces a properly segmented record where findings are attributed to the correct clinical problem.

Semble's Workflow

Semble's workflow proceeds: select the patient, go to the Consultations tab, create a new consultation note, select the consultation type and date, choose a template or free notes, write into sections (adding more via tags as needed), toggle "Show in Summary" for each section, and save. Consultation notes can also be started from an appointment in the calendar. Auto-save protects work in progress.

The flow is straightforward and requires less cognitive overhead. For clinicians who prefer writing narrative notes without structured data entry, this directness is a genuine advantage.

The Resulting Clinical Record

When a Jump consultation is finalised, it produces a multi-layered clinical record: problem-level documentation with SNOMED-coded diagnoses and episode types, structured observations linked to specific problems, coded allergies with reactions and criticality, social and family history panels with coded entries, procedure records with status and body site, immunisation reviews with prior history and clinical decisions, referrals with coded services and priority, medication links documenting prescribing decisions in context, document links connecting letters to the consultation, template snapshots preserving the data capture framework, AI provenance metadata tracking which data was AI-suggested versus manually entered, and a full audit trail with timestamps and user attribution.

Every piece of structured data is linked to both the consultation and the specific problem, enabling queries like "what observations were recorded for this patient's asthma review on this date" - not just "what happened in this consultation." Consultations can be amended after finalisation with a required amendment reason classified as correction, addendum, or clarification. Each amendment preserves a complete before-and-after snapshot of the consultation record, creating an immutable change audit trail while ensuring the original documentation is never lost.

When a Semble consultation is finalised, it produces a consultation note containing the sections the clinician wrote, with Problem Heading and Medical History entries marked Active or Inactive, and allergies surfaced to the patient summary. The note is stored in the Consultations tab and relevant sections appear on the Summary page. The record is a clinical document - readable, shareable, and useful - but not a structured dataset in the same way.

Where Semble Has the Edge

Video consultations. Semble's built-in video consultation platform is feature-rich: virtual waiting rooms, blurred backgrounds, noise cancellation, and support for group calls up to 200 participants. It's tightly integrated with the consultation workflow, allowing clinicians to access patient records alongside the video call. Jump supports remote consultations via delivery mode configuration and meeting URLs, but does not document the same depth of built-in video features.

Consultation diagrams. Semble offers consultation diagrams for annotating anatomical illustrations during consultations. This is useful for surgical, orthopaedic, and dermatological specialties where visual documentation matters. Jump does not document an equivalent feature.

Simplicity and speed. Semble's section-based free text approach has lower cognitive overhead. For clinicians who think in narrative and find structured data entry interruptive, Semble's "just write" model is faster per consultation. Not every practice needs or wants the data richness that POMR provides. A solo practitioner running straightforward self-pay consultations may prefer Semble's directness.

75+ specialty support. Semble markets itself as supporting over 75 medical specialties with customisable templates and workflows. While Jump's template system is configurable for any specialty, Semble's explicit specialty focus may mean more out-of-the-box guidance for non-GP specialties.

The Bottom Line

The consultation module is where the difference between Jump and Semble is most architecturally fundamental. It is not a feature comparison in the usual sense - it is a choice between two different models of what a clinical record should be.

Jump is the stronger choice for practices that need: structured, coded clinical records for audit and governance; SNOMED CT coding throughout the consultation; problem-oriented documentation that separates multiple clinical concerns within a single encounter; queryable observations, allergies, and procedures; guideline-based templates with conditional logic, reusable blocks, and immutability; AI that reads the full clinical record and produces coded structured data; integrated voice transcription with medical terminology; or the ability to answer clinical questions from their data rather than from their documents.

Semble is the stronger choice for practices that need: fast, narrative clinical note-taking with minimal structure; integrated video consultations with advanced features; consultation diagram annotation for procedural or surgical specialties; Heidi AI ambient scribing for hands-free documentation; a simpler consultation workflow with lower cognitive overhead; or broad specialty coverage with customisable question-based templates.

For private GP practices building a scalable, data-driven clinical operation - where clinical governance matters, where CQC readiness requires auditable records, where multi-site consistency demands standardised data capture, and where the long-term value of the clinical record depends on its structure - Jump's problem-oriented, SNOMED-coded consultation model is in a fundamentally different category. It produces a clinical database. Semble produces clinical documents. Both have value, but they serve different futures.

Methodology

This comparison was compiled in February 2026 using Semble's public help centre documentation (help.semble.io), product pages (semble.io), the Heidi AI integration documentation (heidihealth.com), verified user reviews, and direct inspection of Jump EHR's codebase and product functionality. Where a feature is described as "not documented" for either platform, it means we could not find public evidence of the feature - it may exist but not be publicly visible. We encourage readers to verify specific capabilities during a product demonstration.