Units
1. Introduction to Artificial Intelligence in Midwifery: Roles, Boundaries, Validation and Ethics 2. Pregnancy Follow-up and Antenatal Monitoring: History, Appointment and Reminder 3. Risk Early Warning: Preeclampsia, Gestational Diabetes and Danger Signs 4. Mother-Baby Educational Material Production: Brochure, Video Script and Visual Instruction 5. Prenatal Counseling and Empathic Communication: Difficult Conversations 6. Breastfeeding and Postpartum Consultancy 7. Clinical Documentation: Midwife Note, Partograph Summary and Handover 8. Evidence Scanning and Guide Reading: Accessing and Verifying Current Information 9. Privacy, KVKK and Data Security: Protecting Pregnant Data 10. Adapting for Cultural Sensitivity, Multilingualism and Low Literacy 11. End-to-End Workflow, Quality Control and Personal AI Protocol
Unit 2 / 11

Pregnancy Follow-up and Antenatal Monitoring: History, Appointment and Reminder

Gains:

  • Ability to use artificial intelligence as an organizational assistant for story collection, follow-up chart skeleton and appointment reminder texts
  • Ability to distinguish that decisions such as value interpretation, monitoring frequency and screening timing belong to the protocol and the midwife/physician.
  • Ability to verify and customize each message to the pregnant in terms of tone, simplicity and privacy

Pregnancy monitoring — medically known as antenatal monitoring (checking the mother and baby at regular intervals during the prenatal period) — is the backbone of midwifery. A pregnant woman is seen dozens of times from the first application to birth; At each visit, blood pressure, weight, edema, fundus height (the height of the upper border of the uterus measured from the symphysis; gives a rough idea about the baby's growth), baby heart sound and test results are monitored. The quality of this follow-up largely depends on two things: not forgetting any control or finding, and the continuity of communication with the pregnant woman. It is precisely at these two points — regular reminder and story recovery — that AI is a powerful aid.

In this unit, we will cover how to use AI as an organizational assistant that conducts antenatal monitoring without any missed steps. It remains the job of the midwife and physician to interpret the measurement values ​​and make a risk decision; AI collects the story, creates a checklist, and writes clear appointment and preparation messages to the pregnant woman.

Caution: None of the examples in this unit mean "interpret the value, decide if it is normal or abnormal". AI maintains the order of the trace; The midwife determines the meaning of value.

Steps of antenatal monitoring and place of AI

A typical pregnancy tracking consists of the following blocks, and the AI only works on the text/organization side in each block:

  1. Initial application and history taking: Previous pregnancies, chronic diseases, medications, allergies, family history. AI turns the messy notes you enter into an organized story outline.
  2. Appointment schedule: Frequency of checks throughout the trimester. AI writes a calendar and a reminder message to the pregnant woman.
  3. Routine monitoring at each visit: AI produces checklist of items to look at during that visit; midwife signs.
  4. Analysis and screening follow-up: Which analysis was requested in which week and the result came. The AI ​​outputs a tracking table skeleton; The decision on which analysis to request and when belongs to the institutional protocol and the physician.
  5. Summary and homework after the visit: Summary to the pregnant woman: "What did we talk about in this visit, why are you coming to the next visit?" AI writes this in plain language.

Step by step: wrapping up the story with AI

Let's say you have scattered notes from a first application. Follow this order:

  1. De-identify: Remove information such as name, ID, phone number, file number; Write "Pregnant A" instead (details in unit 9).
  2. Paste the raw note: Export your abbreviations and handwriting irregularities as is.
  3. Have it structured: Ask AI for a titled story outline (template below).
  4. Mark missing: Ask AI to list fields that should be asked in the story but are not in the note as "[MISSING]".
  5. Verify and complete: As the midwife, she reads and corrects the draft and completes the missing parts with the pregnant woman.

three mini cases

Case 1 — Missed test (reminder): A midwife is monitoring 90 pregnant women alone in a family health unit. It is noticed that on average 3-4 times a month, the results of an analysis come but no one looks at it. With AI, the midwife creates a framework for each pregnant woman, "Did the expected analysis / result come / was it checked" and prepares a weekly review message template. After three months, the number of "results received but not checked" decreases from 3-4 to 0-1 per month. The AI ​​didn't decide; he merely established the order that ensured that no outcome was overlooked.

Case 2 — Not showing up for an appointment (communication): The rate of not showing up for follow-up appointments in a pregnant group is 22%. The midwife asks AI to draft a reminder message in 3 different tones (short-formal, warm-sincere, very plain for low literacy) and chooses the one suitable for the pregnant woman. After two months, the no-show rate drops to 13%. The difference comes from the fact that the message provides simple and clear information about "when, where, what to bring with you, why it is important."

Case 3 — History collection (time): A newly transferred pregnant woman has scattered notes kept in 3 different institutions. The midwife extracts her identity information and has it summarized by the AI; A 15-minute reading task is reduced to 4 minutes. YZ also points out that there are 2 conflicting dates in the notes; The midwife clarifies these with the pregnant woman.

Copiable templates

Role: You are a documentation assistant working for a midwife.Task: Organize the following raw initial note into the following headings:[Identity: "Pregnant A" only] [Obstetric history] [Medical history] [Medication/allergy][Family history] [Present complaint] [Planned follow-up].Rules: New information DO NOT FAKE. Make a separate list of areas that are not in the note but need to be asked as "[INCOMPLETE - MIDWIFE WILL ASK]". Raw note: [note]

Role: Assistant who drafts an SMS/message to the pregnant woman.Task: Write a follow-up appointment reminder.Information: Date=[..], Time=[..], Location=[..], Bring with you=[..].Produce 3 versions: (1) short-formal, (2) warm-friendly, (3) very simple (low literacy, short words). ADD medical comment/decision.Each version no more than 45 words.

Role: You are the antenatal follow-up organization assistant. Task: "Analysis/scan follow-up TABLE SKELETON" appears for this pregnant woman. Columns: Scan name | Planned week range [FILLED IN BY MIDWIFE/PHYSICIAN]| Requested | Have the results arrived? Did the midwife review it? Rules: YOU decide which scan will be done and when; leave that column blank, the midwife will fill it in according to the protocol.

Role: Post-visit summary assistant. Task: Write a SIMPLE summary to be given to the pregnant woman from the visit note below: "What we talked about today", "What will you do at home", "What is the next appointment for". Rules: Adding decisions/comments; simplify only what is written in the note. Length: 120 words or less, 6th grade level. Visit note: [note]

Weak prompt / Strong prompt

Weak: "Write an appointment message to the pregnant woman."

Result: A generic, unpersonalized, uniform message that perhaps oozes medical commentary.

Strong:

Role: assistant who writes messages to the pregnant woman. 15 August 10:00, X Family Health Center, write a check-up appointment reminder. Ask him to bring his final analysis paper with him. Don't make medical comments. Very simple language, maximum 40 words, one message.

Result: A message that provides clear information about who, when, where and what to bring, whose tone and length are controlled, and which does not leak comments. Difference; is to define the context, border and form from the beginning.

Who decides appointment frequency?

business

Does AI do it?

whose decision

Write an appointment reminder text

Yes (draft)

midwife confirms

Putting the story into headlines

Yes (draft)

Midwife corrects

Assay tracking table skeleton

Yes (empty skeleton)

Midwife/physician fills in

Which test is required in which week?

no

Protocol + physician

Whether a value is normal/abnormal

no

midwife/physician

Decision to increase monitoring frequency

no

midwife/physician

Tip: Read each message out loud to the pregnant woman before sending it. AI sometimes makes innocent-looking but anxiety-provoking sentences, such as "Don't lose control, it's important for your health." Tone is part of your relationship with the pregnant woman; You set it.

Common mistakes

  • Making AI interpret: "My blood pressure is 140/90, what should I do?" The question is not an observation, but a clinical decision; You cannot ask AI.
  • Pasting credentials with story: Forgetting to clear name/ID/phone when giving raw grade.
  • Writing to everyone with a single message tone: Every pregnant woman's literacy and cultural context is different; adapt the tone.
  • Let AI determine the timing of the analysis: The decision about which scan is requested and when belongs to the protocol.
  • Putting YZ's summary in the file without verifying it: The summary does not become an official document until the midwife reads and approves it.
  • Overwhelming the pregnant woman by making reminders too frequent: Continuity of communication does not mean bombardment of messages.

In summary

The real value of artificial intelligence in antenatal follow-up is to establish an organizational discipline that ensures that no control or finding is missed: it collects the history, creates a skeleton of the follow-up chart, writes simple and correct-toned appointment/summary messages to the pregnant woman. Interpretation of the value, monitoring frequency and clinical decision are left to the midwife and physician. Read every message before sending, do not file every story summary without verifying it, never enter identification information into the prompt.

Application task

Take the raw initial reference notes of a pregnant woman you have watched (de-identified) and turn it into a neat draft with the story template; Add the fields that AI marked as "[MISSING]" to the question list for the next visit. Then, create a three-tone appointment message for the same pregnant woman, choose the most appropriate one, and write the reason in one sentence.

checklist

  • [ ] I removed all identifying information before giving the story to the AI.
  • [ ] I read and edited the story draft of YZ; I added the missing questions to the list of questions.
  • [ ] I determined the assay/scan timing based on the protocol, not the AI.
  • [ ] Before sending the message to the pregnant woman, I read it out loud and checked its tone.
  • [ ] I did not leave any value interpretation/clinical decisions to the AI.
  • [ ] I approved the summary as a midwife before putting it in the file.
Previous Next