Community pharmacy · Playbook 74

Patient messaging: New Medicine Service, hypertension case-finding, recalls

The right patients invited at the right time, in words the superintendent's team approved. Every opt-out honoured, every clinical reply answered by a pharmacist.

  • Agents with tools
  • Speech and voice
  • Predictive ML
  • Human in the loop
  • NHS NMS and HCFS specs
  • NHSBSA
  • PECR
  • UK GDPR
  • DCB0129 / DCB0160

At a glance

The problem
Follow-ups and invitations depend on branch teams finding time to phone patients, so windows are missed and eligible people are never asked
Typical volume
Around 40,000 NMS consultations and 12,000 blood pressure checks a month across 400 branches
Who's involved
Superintendent pharmacist, branch teams, clinical governance, the PMR and service IT suppliers, data protection, the patient contact team
Pilot
8–10 weeks in 15–20 branches, after a 3–4 week discovery
Platforms
AWS, Azure, or a sovereign or on-prem stack

An illustrative scenario. The numbers are typical examples, not a client case study. The cloud services and UK rules were checked on 9 November 2026.

The business problem

The New Medicine Service (NMS) runs on fixed windows. A patient starting a new medicine for a listed long-term condition, now including depression for adults, agrees to the service when the first prescription is dispensed. The intervention consultation must happen 7 to 14 days later, and the follow-up 14 to 21 days after that. Only a pharmacist can hold the consultations, in person or by phone or video; text messages can only support them. In a busy branch, someone has to remember who is due, ring them, ring again, and book a slot the pharmacist can keep. When that slips, the window closes and the patient loses the support.

The Hypertension Case-Finding Service, known to patients as the NHS Blood Pressure Check Service, changed on 3 September 2026. Clinic checks are for adults aged 40 and over with no diagnosis of high blood pressure and no check through the service in the last five years, with exceptions recorded. Where a patient misses their ambulatory monitoring appointment, the pharmacy must try to rebook at least twice within seven days and tell the GP on day seven if it can't reach them. The specification lets the whole team promote the service, but finding eligible people and inviting them fairly is manual today.

Messaging patients sounds simple, and it is easy to get wrong. A text that names a medicine or hints at a condition can reach a shared phone. An NHS invitation with a line about a private service becomes marketing. A patient who said 'stop' on one channel is called on another. And at group scale, caps matter: NMS payments are capped at 0.9% of a branch's monthly items from June 2026, so inviting beyond capacity creates work nobody is paid for and appointments the pharmacist can't keep.

What AI changes

  • Patients due an NMS intervention or follow-up get a short text inviting them to pick a time for their call with the pharmacist, sent inside the window, not after it.
  • People who may be eligible for a blood pressure check are found from the group's own dispensing and service records, using the specification's rules, then invited in the order a model suggests, within each branch's capacity.
  • Missed ambulatory monitoring appointments trigger the rebooking attempts the specification requires, and the GP notification on day seven is drafted for the branch to send.
  • Every message is built from templates the superintendent pharmacist's team has approved. The AI chooses the template, the channel and the time, and fills in the date, branch and booking link. It never writes clinical wording of its own.
  • Automated booking calls say at the start that they are automated and offer a person at any point. Replies that ask anything clinical go straight to a pharmacist.
  • One contact-preference register covers every channel, so an opt-out by text, call, app or in branch stops all non-essential messages within minutes.

How success is measured

Agree these, and measure the baseline, before anything is built. Example targets are typical, not promises.

NMS completed within the windows

Share of patients who agreed to the NMS and had both consultations inside the specified windows.

ExampleOver 80%, measured against today's baseline

Blood pressure checks reaching people who need them

Eligible checks completed, split by age band, deprivation and ethnicity where recorded, so outreach doesn't just favour the easiest to reach.

ExampleNo group worse off than today

Clinical replies answered by a pharmacist

Time from a patient's clinical question to a pharmacist's response, measured end to end.

ExampleSame working day

Messages sent against an opt-out

Any message sent to a patient who had opted out of that kind of message, on any channel.

ExampleZero, checked daily

Opt-out and complaint rate

How many patients opt out or complain after a message. Rising numbers mean the timing, frequency or wording needs review.

ExampleStable, and reviewed weekly

Work within caps and capacity

Booked consultations against each branch's NMS cap and pharmacist diary, so invitations don't create unpaid or unkeepable appointments.

ExampleWithin cap in every branch

UK regulation and controls

Cross-sector rules are on the UK regulation map.

  • NHS EnglandNMS service specification and Terms of Service

    Consultations are held by a pharmacist, in person, by phone or by video. Text can only supplement them, so messages book and remind; they never ask clinical questions or replace the consultation. Messages are timed to the 7–14 and 14–21 day windows, and the design records the reason when a window is missed in exceptional circumstances.

  • NHS EnglandHypertension Case-Finding Service specification (revised, in effect from 3 September 2026)

    Eligibility rules come from the specification: aged 40 and over, no hypertension diagnosis, no check through the service in the last five years, with exceptions recorded. Rules, not the model, decide eligibility. Invitations use the patient-facing name, the NHS Blood Pressure Check Service, and never suggest the patient has high blood pressure. Missed ambulatory monitoring appointments follow the rebooking and GP notification steps.

  • NHSBSACaps, claims and post-payment verification

    NMS caps fell to 0.9% of monthly items for activity from June 2026, and new high-level caps for the blood pressure service are still to be agreed. Caps and claim windows are configuration, not code. The messaging platform never creates or submits a claim; the service record stays the evidence.

  • ICOPECR: service messages and marketing

    NHS England's information governance guidance (last updated May 2026) treats invitations that are necessary for delivering a health service, such as screening and vaccination invitations, as service messages, not marketing. It is written for public authorities, and a pharmacy group is an independent contractor, so the DPO confirms the position for each message type. Either way, NHS service messages carry no promotion, and anything about private services goes through a separate, consent-based marketing flow. PECR penalties rose to UK GDPR levels from 5 February 2026.

  • ICOUK GDPR, the common law duty of confidentiality and the national data opt-out

    Contacting patients about NHS services they are receiving or eligible for is direct care: public task and health care conditions (Articles 6(1)(e) and 9(2)(h)), with implied consent under the common law duty where it fits patients' reasonable expectations. The national data opt-out doesn't apply to individual care. Patients are told in the privacy notice, a wish not to receive reminders is respected, texts carry no medicine names or conditions, and a DPIA is done before the pilot.

  • NHS EnglandDCB0129 and DCB0160 clinical safety

    A missed follow-up, a message to the wrong person or a clinical reply that sits unread are clinical hazards. The group needs a clinical safety officer, a hazard log and a safety case before the pilot. The standards are under review (consultation closed 11 September 2026), so expect revised versions.

The solution

Showing AWSAzuresovereign / on-prem services. Switch with the toggle; every diagram and table on this page follows it.

Patient messaging: New Medicine Service, hypertension case-finding, recalls: reference architecture
  1. 01Channels

    • Text message
    • Automated booking call
    • NHS App (once available to pharmacy)
  2. 02Intake

    • API GatewayAPI ManagementKong, or your existing API gatewayDispensing and service events
    • AWS End User Messaging SMSAzure Communication Services SMSA UK SMS gateway through your API layerTexts, replies and opt-outs
    • Amazon Connect with Amazon LexAzure Communication Services Call Automation with Azure Speech in Foundry ToolsOpen-weight speech models (e.g. Whisper) with an open-source telephony platform (e.g. Asterisk)Voice calls
  3. 03Intelligence

    • SageMaker AIAzure Machine LearningMLflow and Kubeflow on KubernetesContact model: who, when, which channel
    • Agent
      Bedrock AgentCoreFoundry Agent ServiceLangGraph on KubernetesInvitation agent
    • Agent
      Bedrock AgentCoreFoundry Agent ServiceLangGraph on KubernetesReply and call assistant
  4. 04Orchestration

    • Step FunctionsLogic Apps / Durable FunctionsTemporal, or n8nInvitation and recall workflow
    • EventBridge + SQSEvent Grid + Service BusApache Kafka or RabbitMQEvents and queues
  5. 05Systems of record

    • Dispensing system (PMR)
    • Assured NMS and blood pressure service IT
    • Contact-preference and opt-out register
  6. 06People

    • Person
      Pharmacist: every clinical reply
    • Person
      Branch team: bookings and call-backs
    • Person
      Superintendent's team: templates
  • Cloud or platform service (AWSAzuresovereign / on-prem)
  • AI agent
  • A person decides
  • Your existing systems
  • Outside your organisation

Components, side by side

Patient messaging: New Medicine Service, hypertension case-finding, recalls: components on AWS, Azure and a sovereign or on-prem stack
LayerAWSAzureSovereign / on-prem
Intake: events from dispensing and service systemsAgree in discovery what your PMR and service IT suppliers can send as events, and what has to be read on a schedule.API Gateway, EventBridgeAPI Management, Event GridKong or your existing API gateway; Apache Kafka
Text messagesAmazon Pinpoint reached end of support on 30 October 2026; new builds use AWS End User Messaging. Use a sender that can receive replies, because a one-way alphanumeric sender can't take a 'STOP' text.AWS End User Messaging SMSAzure Communication Services SMS (UK data location)A UK SMS gateway through your API layer
Automated booking callsNo predictive dialling: Ofcom's persistent misuse policy covers silent and abandoned calls. Test speech recognition on UK accents and older callers.Amazon Connect (London) with Amazon Lex and Amazon PollyAzure Communication Services Call Automation, Azure Speech in Foundry ToolsOpen-weight speech models (e.g. Whisper) with an open-source telephony platform (e.g. Asterisk)
NHS App and NHS NotifyNHS Notify sends to the NHS App first, then text, then letter. Work to bring it to community pharmacy systems is under way but not generally available, so design the channel as a slot to fill, not a dependency.Through your service IT supplier, once offeredThrough your service IT supplier, once offeredThrough your service IT supplier, once offered
Contact modelPredicts the best time and channel and who may need a call from a person. It orders and routes invitations; it never removes an eligible patient from the list.SageMaker AIAzure Machine LearningMLflow and Kubeflow on Kubernetes
Template filling and reply sortingChooses an approved template and fills its slots, and sorts replies into booking, rescheduling, opt-out or clinical. Both clouds offer a choice of models, including Claude.Amazon BedrockMicrosoft FoundryOpen-weight models served with vLLM on your own GPUs, or IONOS AI Model Hub
Agent runtimeThe original Bedrock Agents closed to new customers on 30 July 2026; new builds use AgentCore.Amazon Bedrock AgentCoreFoundry Agent ServiceAn open-source agent framework (e.g. LangGraph) on Kubernetes
WorkflowHolds each patient's journey: window opens, invite, reminder, call, booked, or handed to the branch.AWS Step FunctionsLogic Apps or Durable FunctionsTemporal, or n8n
GuardrailsA final check before sending: the message must match an approved template exactly apart from its slots, with no medicine names, conditions or promotions.Bedrock GuardrailsFoundry guardrails (Azure AI Content Safety)Open-source guardrails (e.g. NeMo Guardrails, Llama Guard)
PlatformManaged AWS servicesManaged Azure servicesKubernetes on-prem (e.g. OpenShift, Rancher), or IONOS Managed Kubernetes
Identity, keys, data protectionMobile numbers are personal data in their own right. Keep them in the register, not in prompts or logs.IAM Identity Center, KMS, MacieEntra ID, Key Vault, PurviewKeycloak or your existing Active Directory; OpenBao or HashiCorp Vault; an HSM
Monitoring and auditLog every message and call: template and version, why the patient was chosen, delivery receipt, reply and who handled it.CloudWatch, CloudTrailAzure Monitor, Log AnalyticsPrometheus, Grafana, OpenTelemetry
UK data residencyAzure Communication Services lets you choose a UK data location; check where calls and messages are processed for your setup.eu-west-2 (London)UK South / UK WestYour own UK data centre, or a UK or EU provider (IONOS AI Model Hub runs in Germany)

On a phone, the table shows the platform you pick. Sovereign / on-prem means open, portable components you can run in your own data centre or with a UK or European provider; IONOS is shown as one example, and others exist. What sovereign cloud means →

Hybrid and on-prem

Many organisations keep core systems and sensitive data on their own servers. This is how the same design works across both.

Stays on-prem

The PMR, the assured service IT and the clinical records stay where they are, with their suppliers or in the group's data centre. So does the contact-preference register if the group already runs one. The messaging platform never becomes the clinical record.

Uses cloud AI

The contact model, template filling, reply sorting and the voice service run in a UK cloud region, or on the group's own GPUs and telephony for the sovereign option. They see only what a message needs: first name, branch, due date and the patient's contact preferences.

How they connect

Through the suppliers' published APIs and an API gateway, over a private link such as AWS Direct Connect or Azure ExpressRoute from the group's network. Bookings go back into the service IT the same way, and opt-outs flow both ways.

If the link drops

No messages are sent, and nothing is sent late: an invitation whose window has passed is dropped and the patient goes on the branch's call list instead. Branch teams work from the due list in the service IT, as they do today.

The agents

What each agent may do through its tools, what it may not, and where a person signs off. The limits are enforced by the platform, not the prompt.

Invitation agent

May do

  • Read the due and eligible lists produced by the specification's rules
  • Take the contact model's suggested order, channel and time, within each branch's capacity and cap
  • Choose an approved template and fill its slots: name, branch, date, booking link
  • Send texts and schedule booking calls through the messaging services
  • Put patients it can't reach on the branch's call-back list

May not do

  • Write or change clinical wording, or use a template that isn't approved
  • Name a medicine or condition, or suggest a diagnosis
  • Contact anyone who has opted out of that kind of message
  • Add promotions or anything about private services
  • Change eligibility rules or decide someone isn't eligible

Where a person signs offThe superintendent pharmacist's team approves every template, every change to one, and the eligibility rules. The agent's permissions are enforced by the platform's identity and policy controls, not by instructions in a prompt.

Reply and call assistant

May do

  • Sort replies into booking, rescheduling, opt-out, clinical or upset
  • Book or move a consultation slot in the branch diary
  • Run an automated booking call that says it is automated and offers a person
  • Record opt-outs in the register straight away
  • Pass clinical or distressed replies to a pharmacist, with the conversation so far

May not do

  • Answer a clinical question, or comment on symptoms, side effects or readings
  • Hold any part of an NMS consultation
  • Keep calling after the agreed number of attempts
  • Ask for a date of birth or other identifiers by text

Where a person signs offA pharmacist answers every clinical reply. Replies that suggest the patient is unwell or at risk go to the pharmacist on duty straight away, with a clear safety-net message to the patient while they wait.

Middleware and integration

The PMR and the assured service IT stay the record. Dispensing a first prescription for a listed condition, a patient agreeing to the NMS, a blood pressure check or a missed monitoring appointment each becomes an event on a queue. Where a supplier can't send events, a scheduled read of the due lists does the same job.

A single contact-preference register sits between every channel and every system. Each message checks it at the moment of sending, not when the list was built, so an opt-out that arrives at 9:01 stops the 9:05 call. Every message carries an ID, so a retry never sends twice.

Bookings and outcomes go back into the service IT through its API, so the pharmacist sees them where they already work and the record supports the claim. The NHS App is added as a channel through the supplier once NHS Notify reaches community pharmacy systems; the workflow doesn't change.

Process and people

  • A template library owned by the superintendent pharmacist's team, with versions, approval dates and a review cycle. A new template goes through the same clinical safety check as a code change.
  • A pharmacist rota for clinical replies, with a target response time and cover for evenings and weekends, or messages that tell patients when replies are read.
  • A daily check that no message went against an opt-out, owned by the patient contact team.
  • A monthly equity review of who was invited and who completed, by age, deprivation and ethnicity where recorded.
  • Branch teams trained before go-live on what the system sends, what it never sends, and how to take a patient's opt-out in person.
  • A clear line between NHS service messages and private-service marketing, with separate owners and separate consent records.

Delivery plan

  1. 3–4 weeks

    Discovery

    Map today's NMS and blood pressure follow-up, measure window completion and opt-outs, agree what the PMR and service IT suppliers can send and receive, and confirm the PECR and lawful-basis position with the DPO.

  2. Gate · must be true to passBaseline agreed, supplier route confirmed, clinical safety officer named, DPIA and hazard log started, templates drafted.

  3. 8–10 weeks

    Pilot

    NMS reminders by text in 15–20 branches. The contact model runs in shadow mode, ranking patients while branch teams call as normal, and the two are compared.

  4. Gate · must be true to passWindow completion up, zero messages against opt-outs, clinical replies answered on time, superintendent sign-off.

  5. 4–6 weeks

    Production

    Text reminders live in the pilot branches, then blood pressure invitations and the monitoring rebooking steps, with monitoring and a rollback plan.

  6. Gate · must be true to passMonitoring, support model, rollback plan, pharmacist reply rota and branch training in place.

  7. Ongoing

    Scale

    Add branches in waves, then automated booking calls, then other recalls the specifications allow, and the NHS App once the supplier route opens.

  8. Gate · must be true to passEach new message type and channel passes the same pilot gate.

Roles and RACI

Who is responsible, accountable, consulted and informed, by activity
ActivitySuperintendent pharmacistProduct ownerSolutions architectI can fill thisTechnical PMI can fill thisEngineersI can fill thisClinical safety officerIG & DPOSecurityI can fill this
Business case and benefitsAccountableResponsibleConsultedConsultedInformedConsultedInformedInformed
Templates and eligibility rulesAccountableConsultedConsultedInformedInformedResponsibleConsultedInformed
Target architecture and supplier integrationConsultedConsultedAccountable and ResponsibleConsultedConsultedConsultedConsultedConsulted
PECR position, DPIA and privacy noticeConsultedConsultedConsultedInformedInformedConsultedAccountable and ResponsibleConsulted
Clinical safety case and hazard logAccountableConsultedConsultedInformedConsultedResponsibleInformedConsulted
Build and integrationInformedConsultedAccountableResponsibleResponsibleConsultedInformedConsulted
Go-live decisionAccountableResponsibleConsultedResponsibleInformedConsultedConsultedConsulted
Run, monitor and improveAccountableResponsibleConsultedConsultedResponsibleConsultedConsultedConsulted

Business case and benefits

Superintendent pharmacist
A · Accountable
Product owner
R · Responsible
Solutions architect
C · Consulted
Technical PM
C · Consulted
Engineers
I · Informed
Clinical safety officer
C · Consulted
IG & DPO
I · Informed
Security
I · Informed

Templates and eligibility rules

Superintendent pharmacist
A · Accountable
Product owner
C · Consulted
Solutions architect
C · Consulted
Technical PM
I · Informed
Engineers
I · Informed
Clinical safety officer
R · Responsible
IG & DPO
C · Consulted
Security
I · Informed

Target architecture and supplier integration

Superintendent pharmacist
C · Consulted
Product owner
C · Consulted
Solutions architect
A/R · Accountable and Responsible
Technical PM
C · Consulted
Engineers
C · Consulted
Clinical safety officer
C · Consulted
IG & DPO
C · Consulted
Security
C · Consulted

PECR position, DPIA and privacy notice

Superintendent pharmacist
C · Consulted
Product owner
C · Consulted
Solutions architect
C · Consulted
Technical PM
I · Informed
Engineers
I · Informed
Clinical safety officer
C · Consulted
IG & DPO
A/R · Accountable and Responsible
Security
C · Consulted

Clinical safety case and hazard log

Superintendent pharmacist
A · Accountable
Product owner
C · Consulted
Solutions architect
C · Consulted
Technical PM
I · Informed
Engineers
C · Consulted
Clinical safety officer
R · Responsible
IG & DPO
I · Informed
Security
C · Consulted

Build and integration

Superintendent pharmacist
I · Informed
Product owner
C · Consulted
Solutions architect
A · Accountable
Technical PM
R · Responsible
Engineers
R · Responsible
Clinical safety officer
C · Consulted
IG & DPO
I · Informed
Security
C · Consulted

Go-live decision

Superintendent pharmacist
A · Accountable
Product owner
R · Responsible
Solutions architect
C · Consulted
Technical PM
R · Responsible
Engineers
I · Informed
Clinical safety officer
C · Consulted
IG & DPO
C · Consulted
Security
C · Consulted

Run, monitor and improve

Superintendent pharmacist
A · Accountable
Product owner
R · Responsible
Solutions architect
C · Consulted
Technical PM
C · Consulted
Engineers
R · Responsible
Clinical safety officer
C · Consulted
IG & DPO
C · Consulted
Security
C · Consulted

R ResponsibleA AccountableC ConsultedI Informed· not involved

Where I'd fit

Option A

Lead the architecture

If there's no architect on the programme, I'd own the design end to end: the events from your PMR and service IT, the contact-preference register, the messaging and voice services, the model and the gates. I'd work closely with your clinical safety officer and DPO, and report to your programme board and superintendent pharmacist.

Option B

Support your architect

If your partner or your own team already has an architect, I'd take on the messaging workflow, the opt-out register and the supplier integration under their lead, and give them a second senior opinion on the model, the voice service and the clinical safety evidence.

Where this goes wrong

  1. Goes wrong

    A text names the new medicine, and the patient's family reads it on a shared phone.

    How the design avoids it

    Templates with no medicine names or conditions, checked by guardrails before every send.

  2. Goes wrong

    An NHS reminder gains a line about the travel clinic, and becomes marketing.

    How the design avoids it

    Service messages and marketing in separate flows, with separate owners, and a template check that blocks promotions.

  3. Goes wrong

    A patient says 'stop' by text and gets a call the next day.

    How the design avoids it

    One register for every channel, checked at the moment of sending, and a daily opt-out audit.

  4. Goes wrong

    A patient replies 'I feel dizzy since starting these' and the reply sits in a queue overnight.

    How the design avoids it

    Clinical replies routed to the pharmacist on duty straight away, with a safety-net message telling the patient what to do if it gets worse.

  5. Goes wrong

    The model learns to invite the people most likely to say yes, and outreach drifts away from those who need it.

    How the design avoids it

    The model orders invitations but never drops an eligible patient, and a monthly equity review checks who was reached.

On the horizon

Changes already announced or under way that affect this design. More in the UK horizon scan.

  1. Under way
    NHS Notify for community pharmacy: work with NHS England and IT suppliers to let pharmacy systems send NHS App messages.

    Build the channel as a slot in the workflow, with the same templates and opt-outs, so adding the NHS App is a configuration change.

  2. During 2026/27
    New high-level caps for the blood pressure service, still to be agreed, and the NMS late-claims window, with its start date still to be confirmed.

    Keep caps and claim windows as configuration, and plan invitations against them per branch.

  3. Late 2026 and 2027
    The review of DCB0129 and DCB0160 (consultation closed 11 September 2026) is expected to bring revised clinical safety standards.

    Keep the hazard log and safety case structured and current, so moving to the new versions is an update, not a rewrite.

  4. 2026–2027
    The ICO's statutory code of practice on AI and automated decision-making, required by regulations in force since May 2026.

    Keep the reason each patient was chosen, and the human handling of every reply, in a form you could show the ICO.

  5. By December 2028
    The Single Patient Record in the NHS App, with the Health Bill as its legal basis, and community pharmacy joined up over time.

    Record bookings and outcomes in the service IT through national standards, so they can flow into a fuller record later.

Questions to ask your partner or vendor

Good answers are specific. Vague answers to any of these are worth following up.

Copy them into your next supplier meeting

  1. Which of our messages are service messages and which are marketing, and who has confirmed that?
  2. What happens to an invitation whose window has passed while the system was down?
  3. How does an opt-out given in branch reach every channel, and how fast?
  4. Who answers a clinical reply at 7pm on a Saturday?
  5. How do we know the model isn't steering invitations away from the people who need them most?
  6. What does it cost to run per thousand patients reached, at our volume, including call minutes?

Planning something like this?

Whether you need someone to lead the design or to support the architect you already have, let's talk it through.

Sam Naseem
Sam NaseemBirmingham, UK · UK & Europe
Available now