SystemQuality Operations & CAPA

Stability Study Management

Stability study management software that draws each chamber shelf by shelf, reads every approved result onto its tray, applies ICH Q1E in the open and sends clients a co-signed trend statement.

Follow one 18-month result to a signed statement ↓
Watch with transcript →
The problem

A pooled straight line promised a shelf life the data never supported

The stability program lives in a LIMS schedule, a chamber spreadsheet, a regression in a spreadsheet and a letter written in Word. When a long-term result lands, the stability lead copies it into the regression, pools the batches because it is quicker, draws a straight line and writes to the client.

Nobody tested whether the batches could be pooled, the degradant curves upward, and the line runs past what the guideline allows. The client files on it, and the regulator cuts it later. Meanwhile a humidity excursion is judged for the whole chamber, so either every tray is questioned or the one capsule tray on the top shelf passes on the average.

Built forQC stability leadsStability coordinatorsQA stability reviewersHeads of QC

The Problem Today

When pooling goes untested and the line runs past the Q1E limit, the shelf life a client files promises more than the data hold.

Four records, joined by hand

  • Pull dates
  • Chambers
  • Regression
  • Letter
  • Pull windows missed quietlyA late pull surfaces as a deviation, often at inspection
  • Excursions judged chamber-wideEvery tray questioned, or the sensitive one missed

The Solution

  1. Every tray on its shelfEach chamber drawn rack by rack, each tray with its next pull and window
  2. Each result read as it landsPlaced on its tray and checked against its own trend first
  3. Pooling tested, Q1E appliedThe shortest batch governs, and the guideline caps how far data extrapolate
  4. Draft attacked, then signedA reviewer agent reads it as the client's regulator would; two people sign

Built-in controls

  • No signature until the sources are opened
  • Any change after signing clears the signatures
  • Each client sees only its own studies

What You Get

Clients file the shelf life their data support, with the evaluation shown so their regulator can follow it.

  • A new analyst works Q1E like a veteranPoolability, model choice and the ceiling worked in view on every draft
  • Out of trend before out of limitEach result checked against its own trend as it lands
  • Clients see their own programSigned statements and next pulls in each client's portal
Features

Stability study management software features, from tray to statement

The QC stability lead reads a client's 18-month result for three tablet batches at 30 °C / 75% RH, while the client hopes to file a longer shelf life.

demoThe CDMO, its chambers, clients, people, products, batches and results are sample data.

  1. 01Chamber map

    Stability chamber map with every tray on its shelf and its next pull

    • Each chamber drawn as it stands. The coordinator finds any tray by rack, shelf and position, with its due or late chip, without the chamber spreadsheet.
    • Scope by client and plant. One client's studies stand out on the map, which is how a CDMO answers a client's question on the spot.
    • Agents at work beside the map. As the result lands, the lead watches each agent read it, trend it and hand it on, with sources.
  2. 02Reviewer

    A regulatory reviewer agent catches an untested pooled regression

    • Read as the client's regulator would. Pooling, the log scale, accelerated data and the pull window are challenged before anyone relies on the draft.
    • The agents' own error corrected. A first pass that pooled three batches without the covariance test was caught and redrawn batch by batch.
  3. 03Q1E evaluation

    ICH Q1E evaluation of stability data, batch by batch, with the ceiling

    • Poolability tested at the 0.25 level. Batches are pooled only when slopes and intercepts allow it, as Q1E sets out, so the shortest batch governs when they differ.
    • The model chosen from the residuals. A degradant that accelerates as the bottle takes up water is fitted on a log scale, with the reason shown.
    • The extrapolation limit worked in the open. Twice the long-term data, but no more than twelve months beyond it, is computed beside the bound, so the client sees why.
  4. 04Person's call

    A dissolution drop the agents leave to the QC stability lead

    • Judgments routed to a person. Whether a drop still inside the limit is reported as a trend is the lead's position, recorded in the statement.
    • The result opened before the decision. The decision stays locked until the lead has read the LIMS record it rests on.
  5. 05Trend statement

    Stability trend statement drafted with a source behind each sentence

    • Study, data, evaluation and chamber history. The client's CMC lead reads the poolability test, the model and the limit in one document.
    • The client owns the shelf-life proposal. The statement says what the data support and leaves the proposal and its commitments to the client.
    • Signed as author with a meaning. The lead signs after opening the result and the covariance table, and the signature says what it means.
  6. 06QA review

    QA signs second after opening the same result and covariance table

    • A second person opens the sources. The QA reviewer reads the records in their own view, so the second signature is a real check.
    • Any later change clears both. A steer, a dropped batch or a new position after signing returns the statement to draft.
  7. 07Client portal

    Each client receives its signed statements in its own portal

    • Only that client's studies. The client sees its studies, next pulls and statements, and never another client's products.
    • Answers in fixed fields. A typed note goes to the stability lead unread by the agents and must be opened before the record closes.
  8. 08What-if

    Shelf life estimation rehearsed before the next time point lands

    • Assumptions shown and editable. The client can see what the next time point would need to show before planning a filing around it.
    • A rehearsal that changes nothing. No study, statement or signature moves, so a hopeful scenario cannot leak into the record.

Coordinators keep the pull windows; the lead and QA each sign the statement

Facilities confirm what happened in a chamber; the client owns the shelf-life proposal.

QC stability leads

Run the evaluations, decide positions and sign each trend statement

  • Poolability, model and Q1E worked in view
  • Their own draft challenged before use
  • Excursions read tray by tray
Stability coordinators

Pull each time point in its window and load the chambers

  • Every tray on a map
  • Windows and late pulls flagged in time
  • Annual commitments tracked by pack
QA stability reviewers

Give the second signature on statements and excursion assessments

  • The same sources the author opened
  • Rule changes that wait for them
  • A sampled check of the week's reads
Heads of QC

Answer for the stability program to clients and inspectors

  • One program across chambers and clients
  • Statements a client's CMC lead can follow line by line
  • A written program that is visibly followed

The trend agent pooled the batches once, and its own reviewer caught it

Each agent reads the stability records without changing them and hands a named person the next step.

Pull scheduler

Keeps each pull inside its window, proposes the year's commitment batches and finds chamber space.

  • Never logs a pull
  • A late pull needs a deviation first
Result reader

Reads each LIMS-approved result and places it on its tray and time point.

  • Never changes a result
  • Reads approved results only
Trend and shelf-life agent

Checks each result against its trend, chooses the model, tests poolability, computes each batch's bound and the Q1E ceiling.

  • Never pools without the covariance test
  • Never proposes a shelf life to a client
Chamber watcher

Reads probes and the mapping study to find which trays an excursion touched, on which shelf and for how long.

  • Never closes an excursion
  • Sends humidity-sensitive packs to a person
Regulatory reviewer

Reads the draft as the client's regulator would and raises each weak point.

  • Never accepts its own point
  • Never decides what a person must
Statement drafter

Drafts the trend statement with a source on each sentence and prepares its transmittal.

  • Never sends unsigned
  • Never states the client's shelf-life proposal

The pooled fit said month 33.5; the shortest batch on its own says 30.6

Pooled when the batches could not be

Slopes differed at p = 0.002 against Q1E's 0.25, so each batch was fitted on its own; the shortest meets 0.50% at month 30.6.

A straight line through curved data

Residuals ran high at months 0 and 18 in all three batches; on a straight line one batch's bound would have reached month 47.

Water above its own trend, still inside the limit

A lyophilized product's water at month 9 read 1.5%, above the range its earlier points predicted and still inside NMT 2.0%; its next statement waits for the laboratory check.

An excursion read tray by tray

Of 25 trays in the chamber during a humidity excursion, 5 in moisture-permeable packs or with capsules went to a person.

demoSample data, from the working solution.

More screens

Five of twenty-five trays in one excursion went to a person

Around the statement: the excursion, the week's results, the pulls, the year's commitments and the rules in force.

An excursion read by shelf probe and packEach tray's exposure from its own shelf probe, the pack's moisture barrier and the dosage form decide which trays a person must assess.
The week's reads, and one out of trendThe week's reads placed on their trays and checked; a water result at 1.5% sits above its expected range and holds the next statement for a laboratory check.
Pulls with their windows and placesThis week's pulls with window, chamber, rack, shelf and tray; a past-window pull is locked until a deviation is raised.
The ongoing stability program, pack by packOne batch a year of each product, strength and primary pack, as EU GMP Chapter 6 says: 20 kept, 3 still to place, each with the scheduler's next step.
Rules a QA signature puts in forcePull windows, the out-of-trend interval and the excursion allowance, each with its source; a change waits for the QA stability reviewer, and its proposer cannot approve it.

demoSample data, from the working solution.

Each shelf's own probe decides which trays an excursion touched

LIMS stability resultsApproved results by batch, condition and time point, with the analyst and reviewer.
Chamber monitoring and mappingProbe readings by shelf, alarms and the mapping study for each chamber.
Protocols and specificationsConditions, time points, tests and limits for each study.
Work ordersRepairs that explain an excursion, from the maintenance system.
Deviations and change controlsLate pulls, excursion records and the change behind each study, from the QMS.
Quality agreementsEach client's turnaround for statements and who owns the shelf-life proposal.

What QC leads ask about stability study management software

What is stability study management software?

Stability study management software runs a stability program: studies, chambers, pulls, results, trends and statements. Here it draws each chamber tray by tray, reads every approved result as it lands, applies ICH Q1E with the working shown, and drafts a statement two people sign before the client receives it.

How does ICH Q1E limit extrapolation?

With no significant change at accelerated and a statistical analysis, the shelf life may reach up to twice the period of long-term data, but no more than twelve months beyond it. Where batches differ, the shortest one governs.

When can stability batches be pooled?

Only when an analysis of covariance at the significance level Q1E names shows slopes and intercepts do not differ. Otherwise each batch is fitted on its own with the pooled error, and the shortest bound decides.

How is an out-of-trend stability result found?

Each new result is compared with the range its own study's earlier points predict, using the interval in the site's procedure. A result outside it goes to a laboratory check even while it is still inside the limit.

Does it propose the shelf life to the client's regulator?

No. The statement says what the data support; the client owns the shelf-life proposal and its filing. The QC stability lead and the QA stability reviewer sign before anything is sent.

How are chamber excursions assessed?

Tray by tray, from the probe on each shelf, the time above tolerance and how well the pack keeps moisture out. Trays in sensitive packs go to the stability lead, and a facilities engineer confirms the cause before QA signs.

What happens to a stability pull that misses its window?

It cannot be logged as on time. The pull schedule shows each tray's window and its place in the chamber, and once the window has passed the pull stays locked until someone raises a deviation. The pull scheduler never logs a pull itself.

Send us one study
and its next time point.

Send a stability study's results and protocol, and we will test pooling, apply Q1E and draft the statement for your review.