<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by KONGDY Health on Medium]]></title>
        <description><![CDATA[Stories by KONGDY Health on Medium]]></description>
        <link>https://medium.com/@kongdyhealth?source=rss-d964862eb840------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*IjJsTzt80iwTC7T4FRmkhg.png</url>
            <title>Stories by KONGDY Health on Medium</title>
            <link>https://medium.com/@kongdyhealth?source=rss-d964862eb840------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:33:12 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@kongdyhealth/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Why Your Manufacturer's Electronic Batch Record Is a Filing Cabinet With a Database Behind It…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-manufacturers-electronic-batch-record-is-a-filing-cabinet-with-a-database-behind-it-0add7b2af51e?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/0add7b2af51e</guid>
            <category><![CDATA[data-integrity]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[electronic-batch-record]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[quality-management]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Thu, 19 Nov 2026 06:07:05 GMT</pubDate>
            <atom:updated>2026-11-19T06:07:05.461Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Manufacturer&#39;s Electronic Batch Record Is a Filing Cabinet With a Database Behind It: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>A paperless batch record or a manufacturing execution system is usually introduced to a buyer as a digital version of the paper process it replaced, and that framing is the source of most of the problems that follow. A record is not a document. It is a set of decisions, each made by a named person, each supported by evidence, each retained for a defined period and each capable of being reconstructed months later when a batch is investigated, a regulator asks or a customer audits. The software does not make those decisions; it either supports them, or it records them badly. The seven specifications below are what a buyer should require before agreeing that a manufacturer can run your product on a system you cannot see.</p><p>Specification 1: The system is justified by the decisions it enables, not by the paper it replaces. Ask what the system actually decides. In a good implementation it closes a batch only when every specified check has a recorded result and every deviation has been dispositioned, it prevents a material from being issued to a batch without a released incoming inspection record, it prevents two operators from working the same operation simultaneously, it captures the equipment state and the line settings at the moment of the event rather than afterwards, and it produces the record that a release decision rests on. Ask also what still requires a person, and which of those people have the authority to override the system and under what record.</p><p>Specification 2: The system is validated, and its validation state is current. Ask for the risk-based classification of the application and of the infrastructure it depends on, the installation, operational and performance qualification evidence, the periodic review cycle, and the change control that covers both version upgrades and configuration changes. The second half matters more than the first. A system validated in 2023, upgraded in 2024 and then reconfigured to accommodate a new line in 2025, with no requalification of the changed configuration, is not a validated system, and an inspector will treat it that way regardless of what the original validation report says.</p><p>Specification 3: The record content and the review workflow are specified, including what each signature means. Ask what is captured per batch: material lots and quantities with the weighing or dispensing step, in-process results with the operator and the equipment used, critical process parameters as they were recorded, deviations and their dispositions, equipment status, and the identity of the person who performed each step. Ask how the review is staged, who reviews which part, at what point in the process the review occurs, what a signature asserts, and what the system blocks until the gate is satisfied.</p><p>Specification 4: Electronic signatures are attributable and cannot be borrowed. Ask how identity is established and how the signature is linked to the record it applies to, what the meaning of each signature type is, whether any step requires re-authentication rather than an open session, how the signature is manifested when the record is exported or printed, and what prevents one person from signing on behalf of another. Ask specifically how shift workers authenticate: if the practical answer is that the login is slow and night shift uses a shared or supervisor account, then the attribution requirement is being met in form and broken in substance, and that is a finding waiting to happen.</p><p>Specification 5: The audit trail is complete, and the data survives. Ask whether every change to a record is captured with the old value, the new value, the identity, the timestamp and the reason, whether anyone can silently alter or delete a recorded value, how long the records and their audit trails are retained and how they are retrieved after that period, whether backup and restore have actually been tested rather than assumed, and whether the audit trail is reviewed or merely collected. An audit trail that is generated and never examined is evidence that a system was bought, not evidence that a process is controlled.</p><p>Specification 6: Interfaces and the master record are defined, and transcriptions are visible. Ask which system is the master for each piece of information, which data moves between the manufacturing execution system, the electronic batch record, the laboratory system, the quality system and the enterprise system, how those interfaces are validated, and what happens to a value that is entered once and used in several places. Ask specifically how manual entries and transcriptions are identified in the record, because a value that was typed by a person rather than captured from an instrument has a different evidential weight, and a reader of the record should be able to tell which is which.</p><p>Specification 7: The buyer’s access is contractual and the operation is resourced. Ask how the buyer obtains a batch record after the fact: within what time, in what form, for which batches, with what assistance, and at what cost. Ask what happens if the system is unavailable, if a supplier changes the platform, or if the manufacturer’s own use of the data for release trending or capacity planning is what the buyer’s evidence actually depends on. A buyer whose access to its own product’s record depends on a discretionary favour from the manufacturer has not been given a system, and the first time it matters will be the worst possible moment.</p><p>The practical test is simple: pick a batch from three years ago, ask for its record, and see what comes back and how long it takes. If it arrives complete, with the materials, the in-process results, the parameters as they were, the deviations and their dispositions, the signatures attributable to named people, and the history of every change, then the system is doing the job a record is supposed to do. If it arrives as a PDF that took three requests and a scheduled export, then the organisation has digitised the filing of its paper and left the substance of the record where it was.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0add7b2af51e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Regulatory Submission Written With AI Will Fail the First Time an Inspector Reads the…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-regulatory-submission-written-with-ai-will-fail-the-first-time-an-inspector-reads-the-79125faddebd?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/79125faddebd</guid>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[ai-regulatory-filing]]></category>
            <category><![CDATA[data-integrity]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[quality-management]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Wed, 18 Nov 2026 06:43:11 GMT</pubDate>
            <atom:updated>2026-11-18T06:43:11.862Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Regulatory Submission Written With AI Will Fail the First Time an Inspector Reads the Manufacturing Section: Seven Specifications Every Buyer Should Require</h3><p>Every regulatory organisation has now had the conversation about artificial intelligence, and the conversation has settled into two positions that are both wrong. The first is that artificial intelligence cannot be used in a submission because it is not deterministic, and the second is that it can be used anywhere because a human reviews it. Both are wrong, and the cost of believing them lands on the company that has to defend the dossier in an inspection. In a transdermal programme the exposure is not abstract: a submission contains statements about manufacturing sites, process parameters, analytical results, stability data and shelf life, and every one of those statements has to be true, traceable and attributable to a person who stands behind it. The seven specifications below are what a buyer should require before any part of a regulatory submission is drafted, assembled or checked with a model.</p><p>Specification 1: The permitted scope is written down per dossier component, and it excludes scientific conclusion. Ask which parts of the submission the organisation has allowed a model to touch, and which it has not. The defensible uses are structural: building a table of contents, assembling and numbering sections, checking cross-references and hyperlinks, checking that tables and figures carry consistent units and footnotes, checking that a document set matches the previous version, converting legacy documents into a structured outline, and flagging stylistic inconsistency across a set of reports. The undefensible use is authoring a scientific conclusion, a stability conclusion, a clinical interpretation or a statement about what the data shows, because those are judgements made by qualified people and the regulator will hold them to that standard.</p><p>Specification 2: Every generated statement has a named reviewer and an accountable approver, and no generated text is submitted unreviewed. Ask what the review record contains, who signs, and what the reviewer is accountable for. Ask specifically what a reviewer is required to check, because a review defined as reading for grammar and consistency is not a review of generated content. Ask whether the reviewing capacity has been increased to match the additional content a model produces, and what happens to a document section where the reviewer is asked to confirm a statement that no one in the organisation has direct knowledge of, such as a process parameter or an analytical result.</p><p>Specification 3: The tool is qualified as a tool, with a risk-based classification, version control and revalidation triggers. Ask how the software has been classified for risk, what its intended use statement is, which versions of the model and of the vendor platform are permitted, what triggers a requalification when the vendor releases a change, and what test cases exist with their expected outputs and the evidence that they pass. A generative component whose behaviour can change without notice is a change to the validated state of the process that uses it, and it has to be managed as one, with the same expectation of documented evidence and a controlled change process that applies to any other piece of validated software.</p><p>Specification 4: The audit trail is complete enough to reconstruct how every output was produced. Ask what is retained for each generated passage: the prompt, the input material, the output, the model and platform version identifiers, the timestamp, and the reviewer’s edits with their identity. Ask whether that record is retrievable later for a specific document and a specific section, and whether it satisfies the same attribution, legibility, contemporaneity, completeness and accuracy expectations that apply to the rest of the record. A model that writes directly into the source document without a retrievable trail of how the text was produced makes that document unverifiable.</p><p>Specification 5: Confidentiality and data-handling terms are contractual, not assumed. Ask the questions that matter commercially before a single document is uploaded: whether submitted content is used to train the vendor’s models, whether any retention or any human review of the content occurs, where the data is processed and stored, whether it is transferred outside the jurisdiction where the submission is filed, and what happens to it on termination. Require the answer in the contract rather than in the vendor’s marketing material, and require the same commitment from any sub-processor the vendor uses. Unpublished manufacturing and clinical information is protected by regulation, and an unassessed tool that learns from it creates an exposure that no regulatory pathway in the dossier will cover.</p><p>Specification 6: The review burden is resourced and the risk is not shifted onto the reviewer, with adversarial testing on the real submission content. Ask what proportion of additional effort the model is expected to create, who is absorbing it, and how the organisation has tested the tool against content it actually cares about, rather than against demonstrations. The useful test is deliberately adversarial: give the tool the real material and look for what it confidently produces that is wrong, unsupported, out-of-context or sourced from a different product. Ask what the results of that exercise were, what was changed as a result, and what the acceptance criteria for the tool are in terms of the error classes it is allowed to make.</p><p>Specification 7: Audit readiness and the rationale are documented, and the tool’s qualification package is available for inspection. Ask whether an inspector or auditor could be shown the tool’s intended use, its risk classification, its test evidence, its version history and its change log, alongside the submission itself. Ask whether the decision to use the tool in a particular submission was recorded with its rationale, and who approved it. This is the point at which a well-run programme separates itself from an organisation that adopted a tool and has not yet worked out what it is accountable for, and it is far cheaper to have the answer ready than to reconstruct it during an inspection.</p><p>The framing that holds up in practice is that a model is a formatting and consistency engine, not an author, and that the qualification, the audit trail and the review effort belong to the organisation that submits the dossier rather than to the vendor that sells the tool. Ask what the tool does, what it is qualified to do, who reviewed what it produced, what was retained, what the vendor is contractually forbidden from doing with the content, and what an inspector would see. If all six answers are clear and in writing, the use of a model is a productivity decision with a control framework around it. If the answers are that the tool is accurate, that the reviewer checks everything, and that the vendor’s terms are standard, then the organisation is relying on three assumptions that none of them supports.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=79125faddebd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Manufacturer's Second Plant Is Not a Backup Site Until You Can Qualify a Lot From It…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-manufacturers-second-plant-is-not-a-backup-site-until-you-can-qualify-a-lot-from-it-19663fb8cd52?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/19663fb8cd52</guid>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[second-manufacturing-site]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Tue, 17 Nov 2026 06:20:16 GMT</pubDate>
            <atom:updated>2026-11-17T06:20:16.918Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Manufacturer&#39;s Second Plant Is Not a Backup Site Until You Can Qualify a Lot From It: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>A second manufacturing site is usually presented to a buyer as an unqualified good: it removes the single-site dependency, it adds capacity, and it protects the supply when one location has a problem. In a transdermal programme the same decision is a quality event before it is a supply event, because a patch made in one building is not interchangeable with a patch made in another until somebody has shown it. The two sites have different raw material lots, different operators, different equipment histories and different stability data, and the regulatory record treats the manufacturing site as part of the product’s identity in most markets. The seven specifications below are what a buyer should require before agreeing to let any of a manufacturer’s volume move.</p><p>Specification 1: The second site is qualified on its own terms, with a site-specific validation package and a date. Qualification does not travel with the company. Ask which qualifications the second site holds in its own name: installation, operational and performance qualification on its own equipment, its own process validation, its own cleaning and change-control records, and its own stability programme covering the products it will make. Ask for the date the site was qualified and which products were in that scope. A site described as validated because the company is validated is not a validated site.</p><p>Specification 2: The product allocation between sites is defined before anything moves, including the identity of a lot. Ask which products are made where, from when, and whether the allocation changes with demand. Ask how lot identity is handled when the site changes: whether lot numbers encode the site, whether the artwork and packaging carry a site identifier, and whether a distributor can tell from the carton which site made the product. Traceability that cannot distinguish the sites is traceability that will fail during a recall, and that is precisely when it is needed.</p><p>Specification 3: Technology transfer is treated as a change event rather than a copy operation. Ask what is actually re-established at the new site: the coating parameters and their ranges rather than the nominal set points, the in-process controls and their sampling plans, the analytical methods and whether they are verified or merely transferred, the cleaning and changeover performance, and the raw material suppliers, whose grades are not automatically equivalent between sites. Ask what evidence will be generated before commercial production, and what the disposition of the first batches is. Copying a recipe onto different equipment is a transfer of paperwork, not of process.</p><p>Specification 4: The regulatory consequence of the site change is identified per market before the transfer starts. In most jurisdictions the manufacturing site is named in the filing or registration, so a change of site is not always a notification and is not always free. Ask which markets are affected, whether the change is a notification, a supplement or a new filing in each one, the timelines and fees, the labelling and artwork consequences, and what the plan is if any market does not accept the change. A transfer that is complete operationally and blocked in one market leaves the buyer with mixed-site stock it cannot sell.</p><p>Specification 5: The mixed-supply period is controlled by a written rule. There is a window in which both sites make the same product, and that window needs a definition. Ask which site supplies which market during it, whether a market may receive lots from more than one site, whether the buyer is notified of every site change for a product, and what happens if the transfer is partial, delayed or reversed. Ask also what the comparability evidence is, because a buyer is entitled to know whether a lot from the second site is expected to match a lot from the first in release, adhesion and stability, and what evidence supports that expectation.</p><p>Specification 6: The quality agreement, change notification duty and audit rights extend to the second site, including its own sub-suppliers. Ask whether the second site appears as a covered manufacturing site in the quality agreement, whether the change notification duty applies to changes made there, whether the buyer’s audit rights extend to it, and whether the access to batch records, deviations and release testing data follows. Ask what is flowed down to the second site’s own suppliers, because a second site frequently carries a different tier of sub-suppliers, and a contract that stops at the manufacturer’s doorstep no longer describes where the product is actually made.</p><p>Specification 7: The continuity benefit is quantified and its cost is accepted explicitly, with the risk allocation written down. A second site removes a single-point-of-failure dependency and it adds split-batch complexity, transfer validation cost, duplicated quality overhead and coordination across two operations. Ask what the manufacturer’s own risk model says, what the disruption scenario was that justified the investment, and which party carries the cost when the two sites produce different quality. The commercially sound arrangement is one where the buyer can see the continuity benefit and the split-batch risk in the same document.</p><p>The practical test of a second-site arrangement is simple: ask whether a lot from the second site would be accepted by the buyer’s own market authorisation, by the buyer’s own quality system and by the buy own customers, without a special argument. If the answer is yes, with the paperwork that supports it, the second site is real capacity rather than a second address on the certificate. If the answer is that it is broadly equivalent and the details will follow, then the buyer has been given a supply option and not a qualified site, and the difference between those two things is discovered during the first inspection.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=19663fb8cd52" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Safety Stock Is a Liability Wearing a Security Uniform: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-safety-stock-is-a-liability-wearing-a-security-uniform-seven-a5290dba92bd?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/a5290dba92bd</guid>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[inventory-management]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 03:38:27 GMT</pubDate>
            <atom:updated>2026-11-16T03:38:27.022Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Safety Stock Is a Liability Wearing a Security Uniform: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>Capacity that cannot be flexed has to be buffered instead, and inventory is the buffer most buyers reach for first. It is also the buffer that goes wrong quietly, because a stock position looks like security on the dashboard and behaves like a liability on the balance sheet. Safety stock in a transdermal programme is not the same kind of thing as safety stock in a general manufacturing business: it occupies validated storage, it has an expiry date, and the buffer that protects against a supply interruption can become the reason a product cannot be launched. The seven specifications below are what a buyer should require before contract signing, because inventory is a decision about how the supply chain behaves, not a number to be set once and left.</p><p>Specification 1: Inventory is classified by purpose, and each class carries its own rule and its own cost. Ask how the manufacturer splits the stock position into the categories that actually behave differently: stock that supports the normal production cycle, stock held to cover demand and lead-time variability, stock that decouples a constrained step from the rest of the line, and stock that is no longer moving. Ask what each category is sized against and what happens when a category is consumed below its intended level. A manufacturer who reports one total inventory number has hidden the most useful information behind the least useful figure.</p><p>Specification 2: Safety stock is derived from measured variability and the real lead time, and the derivation is written down. The inputs should be named: how demand variability was measured, over what history, how lead-time variability was measured, what service level the calculation is intended to deliver, and what assumptions were used for the units that sit between the two. Ask to see the working rather than the result, and ask what happens to the buffer when the input distribution changes, because a buffer sized on last year’s variability and never revisited is a historical artefact.</p><p>Specification 3: The cost of holding is the full cost, not the finance rate. Holding stock in a transdermal programme costs capital, but it also costs validated and temperature-controlled storage, capacity in a stability programme that cannot then be used for a new product, insurance, handling and counting labour, and the paperwork of every batch movement and status change. Ask what the manufacturer charges for storage and whether it is itemised, and ask what the physical space occupied by the buffer displaces. A buffer that sits in the stability chamber is not a free buffer.</p><p>Specification 4: Expiry and shelf life are part of the inventory policy, not a by-product of it. A safety stock that is sized in units rather than in time will, at some point, hold product that is older than its own shelf life, and the older it sits the less likely it is to be used even before it expires. Ask what the expiry profile of the buffer looks like, how old the oldest stock is, what the write-down policy is, and whether the consumption plan is rotation-based so that the stock that was bought first is the stock that moves first. Ask also what happens to a buffer when the underlying forecast changes and the buffer is no longer the right size.</p><p>Specification 5: Replenishment is triggered by a documented reorder point, and the trigger is compatible with the supplier’s real terms. The reorder point has to be consistent with the actual lead time, and the order it produces has to clear the supplier’s minimum order quantity and fit their production slot, because an order that is below their minimum and therefore not manufacturable is not a replenishment plan. Ask what the reorder point is, how it is calculated, who is authorised to change it, and what the consequence is of a reorder that arrives when the supplier has no capacity. Ask also how the system behaves when the forecast and the inventory record disagree.</p><p>Specification 6: The demand signal that sizes the buffer is the same signal the forecast uses, and the two are reconciled. It is common for an inventory target to be computed against one view of demand and for replenishment to be triggered against another, and the gap between them is invisible until it is expensive. Ask which demand signal sizes the buffer, which signal triggers the reorder, whether they are the same series, and how often they are reconciled. Ask also how a deliberate promotion, a regulatory change in one market, or the launch of a second product is reflected, since those are exactly the events that turn a reasonable buffer into an unreasonable one.</p><p>Specification 7: Inventory visibility and control are specified, including ageing, status and the buyer’s right to see it. Ask for the inventory record structure: batch and lot, expiry, location, status including any quarantine or hold state, and the date the stock was received. Ask how often physical counts are performed and how discrepancies are reconciled, what the escalation is when a projected stockout crosses a defined threshold, and what the buyer is entitled to see, at what frequency and in what form. A buffer the buyer cannot see is not a buffer the buyer can rely on.</p><p>The practical test is simple: ask what the buyer would receive in the week the buffer exists to cover. If the answer is a number derived from measured variability, a documented reorder point and an expiry profile that rotates, the arrangement is real. If the answer is a target set in a meeting and reviewed only when it becomes a problem, the buffer is a guess with an expiry date, and it will cost the buyer twice: once in the write-off, and once in the stability capacity the new product could not have.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a5290dba92bd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Capacity Promise Is a Nominal Rate and Not a Delivery Plan: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-capacity-promise-is-a-nominal-rate-and-not-a-delivery-plan-seven-41da4df215f9?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/41da4df215f9</guid>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[capacity-planning]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Fri, 13 Nov 2026 01:33:04 GMT</pubDate>
            <atom:updated>2026-11-13T01:33:04.363Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Capacity Promise Is a Nominal Rate and Not a Delivery Plan: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>Capacity in a transdermal programme is usually quoted as a number of units per year, and that number is almost always a piece of arithmetic performed on the nominal rate of a coating line rather than a measurement of what the line can actually deliver. The gap between the two is where the commercial relationship breaks: the buyer plans against the nominal figure, the manufacturer plans against the achievable one, and the difference appears as late delivery in the first year of a new product. Volatility makes the same gap dangerous rather than merely annoying, because a supplier who is already running near its real ceiling has no ability to absorb a spike and no willingness to hold stock against one. The seven specifications below are what a buyer should require before contract signing, because capacity and volatility are a single commercial subject rather than two.</p><p>Specification 1: The capacity figure is expressed as qualified output, not as equipment rate. Ask what the figure includes and what it excludes. Real deliverable capacity is the nominal rate multiplied by the yield, reduced by the changeover time, the cleaning and validation time between products, the first-article and in-process sampling, and the time the batch records spend in review and release before the goods can be shipped. Ask for the measured rate on the specific products in the range, because the mix determines the number, and a capacity figure quoted for the most efficient product in the range describes a condition that may never recur.</p><p>Specification 2: The ramp-up curve is stated as a schedule of deliverable quantity by period, not as a start date. Equipment that is installed and qualified does not produce on day one. The first batches carry elevated yield loss, repeated first-article checks, engineering trials, stability commitments that consume slots and a documentation burden that slows release. Ask for the first twelve months of committed quantity, the yield assumptions behind it and what gates each increase in rate. A supplier who offers a start date rather than a quantity schedule is offering an aspiration.</p><p>Specification 3: The minimum economic batch is stated, and the volume below which production stops being viable is named. There is a batch size below which a coating run, a die set, a material minimum order and a validation commitment make the product uneconomic to make at all, and below a certain annual volume the changeover consumes the margin entirely. Ask for the minimum batch per product, the minimum annual volume per product, what the changeover cost is per changeover, and what happens to a product that falls below the threshold. Ask also whether the threshold is a floor the manufacturer may enforce, because a buyer who has no answer has bought a product whose supply depends on somebody else’s commercial appetite.</p><p>Specification 4: Demand variation is separated into the part that can be forecast and the part that cannot. A planned ramp for a new product, a seasonal pattern and a promotional peak are forecastable and should be planned with the supplier; a competitor withdrawal, a tender outcome, a regulatory decision in another market and a customer’s own inventory cycle are not, and treating them the same produces either excess stock or a stockout. Ask how the supplier classifies a demand change, what notice the buyer must give and what the response is, and whether the response differs by the type of change. A supplier who has one answer for all variation has not thought about it.</p><p>Specification 5: The split between inventory and capacity as the shock absorber is explicit, and each has a cost. Some of the fluctuation is absorbed by holding stock and some by running the line harder, and the two cost different amounts of money and carry different risks. Ask which mechanism covers which part of the range, how much finished goods inventory is held and what it costs, what shelf life that inventory consumes and what write-down exposure it creates, and what the line can absorb beyond its normal schedule without moving the other commitments. A supplier who absorbs everything through inventory has quietly bought the buyer a working capital problem.</p><p>Specification 6: The commercial terms are matched to the volatility rather than applied uniformly. A minimum volume commitment protects the manufacturer against under-use of a dedicated asset and destroys a buyer whose demand is genuinely variable. Ask what the commitment is, whether it is a take-or-pay or a capacity reservation with a different profile, what happens when the demand is below it, and whether the price mechanism reflects the flexibility the buyer is paying for. Ask also how a change in the product mix, which changes the effective capacity, is handled in the pricing.</p><p>Specification 7: The communication rhythm for volume changes is agreed, with a lead time the customer can plan against. The customer’s own inventory decision depends on how much notice it has. Ask what notice is required for a volume increase and for a decrease, how quickly the supplier can confirm, what the customer is told about capacity availability and any change in it, and what the escalation route is when a commitment is missed. A supplier that confirms changes only after the customer has placed the order has converted a forecasting exercise into a guessing exercise on the customer side.</p><p>Capacity and volatility are two sides of one question, which is whether the supply arrangement can deliver what the market will ask for, when the market asks, at a price that still works. Ask for the qualified output figure, the ramp curve, the minimum economic batch, the split between forecastable and unforecastable variation, the division of labour between inventory and capacity, the matching commercial terms and the communication rhythm, and the arrangement stops being a nominal rate with a caveat. The commercial question worth asking is not whether the line can make ten million units a year, but whether the buyer can rely on nine of them in a bad quarter and can afford the tenth.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=41da4df215f9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Continuity Plan Does Not Cover the Scenario That Ends Your Supply: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-continuity-plan-does-not-cover-the-scenario-that-ends-your-supply-seven-a0efbde1ba09?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/a0efbde1ba09</guid>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[business-continuity]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Thu, 12 Nov 2026 02:08:34 GMT</pubDate>
            <atom:updated>2026-11-12T02:08:34.113Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Continuity Plan Does Not Cover the Scenario That Ends Your Supply: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>A business continuity plan for a transdermal manufacturer is usually written as a document about the factory, and it is usually written for the people who work in the factory. That produces a plan that is excellent at describing what happens to the coating line and silent about what happens to the patients who take the product. The gap matters because the interruption that stops a plant for two days is rarely the interruption that ends a supply chain: the one that matters is the raw material that does not arrive, the pouch specification that cannot be qualified in time, the second site that cannot take the volume, or the customer who needs a delivery date that no scenario can produce. The seven specifications below are what a buyer should require before contract signing, because continuity is a property of a supply arrangement rather than of a building.</p><p>Specification 1: An interruption is defined objectively, and the definition excludes the interruptions that do not matter. A plan that treats every deviation as an interruption will either trigger constantly or be ignored. Ask what constitutes an interruption in measurable terms, such as a failure to deliver a committed quantity within a committed window, and which deviations are explicitly excluded. Ask also whether the trigger is defined from the buyer’s position rather than the manufacturer’s, because the same event is an inconvenience for the plant and an interruption for the customer.</p><p>Specification 2: Recovery objectives are set per scenario rather than globally. A single recovery time objective applied to everything produces a plan that is either impossible or worthless. Ask for the objective in each scenario: the maximum time the buyer can be without supply before switching, the maximum quantity that can be lost before a market notices, and the maximum period for which an alternative source can be used before a regulatory action becomes necessary. Ask also who is allowed to accept a lower objective on a given day, because an objective nobody can waive is not an objective.</p><p>Specification 3: The scenario list is ranked by patient impact rather than by plant impact. The scenarios that belong at the top of a transdermal plan are not the ones that stop the most machinery. They are the loss of a sole-qualified raw material, the loss of the pouch specification, the loss of the only site qualified for a given market, and the loss of a material with a long qualification lead time. Ask for the scenario list with the ranking and the reasoning, and ask which scenarios have been rehearsed. A plan that ranks a coating line failure above a sole-source excipient failure has been written by the wrong department.</p><p>Specification 4: Each scenario has a documented alternative with its prerequisites stated in advance. An alternative that requires a qualification, a change notification, a new site or a customer approval is not an alternative unless the steps are known and the time they take has been measured. Ask for each scenario what the alternative is, what has to be in place before it can be used, how long it takes, and what it costs. Ask specifically which alternatives were never actually tried, because the alternative that has been described many times and never executed is the one that fails during a real event.</p><p>Specification 5: Safety stock for materials and packaging is sized against real lead times and real minimum order quantities. The stock that protects a programme is not the stock that covers the average demand variation; it is the stock that covers the longest credible interruption plus the time to implement the alternative, adjusted for the fact that a small order of a critical material may be below the minimum order quantity and therefore not orderable at all. Ask for the calculation per critical material, the resulting stock value, the shelf life it consumes and the write-down risk it creates, and who owns the decision to release it.</p><p>Specification 6: The communication and notification path is defined for each stakeholder, with a clock. A customer who learns of a supply problem from a delivery that did not arrive has already lost the information that would have let them act. Ask what the buyer is told, when, by whom and with what content, including the expected recovery date and the alternative offered. Ask the same for the regulator, for the market authorisations that depend on the site, and for the employees on the line, whose behaviour during a crisis is part of the response.</p><p>Specification 7: The plan is exercised, and the exercises produce changes to the plan. A continuity plan that has never been run is a document, and the only way to know which parts are real is to rehearse. Ask how often the plan is exercised, whether the exercise is a tabletop or a real switch of material or site, who participates, what was found on the last one and what changed as a result, and when the plan was last revised. Ask also who owns it, because a plan owned jointly by two functions is a plan owned by neither.</p><p>Business continuity for a transdermal product is a supply question wearing an operational costume, and the buyer is the party with the most to lose when it is treated as one. Ask for the objective definition of an interruption, the recovery objectives per scenario, the patient-ranked scenario list, the documented alternatives with their prerequisites, the safety stock sized against real lead times, the communication clock for each stakeholder, and the exercise history, and continuity stops being a document and becomes a capability. The scenario that decides whether a programme survives a bad year is almost never the one the plant prepared for, and the preparation that helps is the one that names the material with no second source and the week the pouch specification changed.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a0efbde1ba09" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Real-World Evidence Will Not Be Accepted Where You Need It: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-real-world-evidence-will-not-be-accepted-where-you-need-it-seven-a450abc1565c?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/a450abc1565c</guid>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[regulatory-affairs]]></category>
            <category><![CDATA[real-world-evidence]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Wed, 11 Nov 2026 02:17:45 GMT</pubDate>
            <atom:updated>2026-11-11T02:17:45.978Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Real-World Evidence Will Not Be Accepted Where You Need It: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>Real-world evidence is the most oversold and under-specified tool in a transdermal programme, because it is usually adopted as a hope rather than as a study design. A registry is set up, data accumulate for three years, and at the end a regulatory question is asked that the data cannot answer, because the question was not the one the data were designed to address. The cost of that sequence is not the registry; it is the three years, the relationship with the sites and the belief that the evidence existed. The seven specifications below are what a buyer should require before contract signing, because real-world evidence only works when the question is fixed before the data are collected, and that ordering is the whole discipline.</p><p>Specification 1: The intended use of the evidence is stated, including what it cannot support. Real-world evidence can support a safety characterisation, a utilisation description, a subgroup hypothesis, a labelling discussion and, in defined circumstances, an extension of indication; it cannot substitute for the randomised evidence that established efficacy. Ask what specific decision the evidence is meant to inform, which market’s decision it is meant to inform, and what the fallback is if the evidence turns out to be supportive rather than decisive. A programme that expects a registry to deliver an indication has misunderstood the instrument.</p><p>Specification 2: The data sources are named, with their selection bias and missingness stated. Electronic health records capture what the clinician recorded, claims data capture what was billed, registries capture what the enrolling site chose to enter, and wearables capture what the device recorded and the patient chose to synchronise. Each has a different missingness mechanism and each selects a different population. Ask which sources are used, what population each represents, what is systematically absent from each, and how the combination is justified. Two sources combined do not cancel each other’s bias; they usually compound it.</p><p>Specification 3: The variables that a transdermal patch can actually be measured on are specified, because they are the ones only real-world data can see. A patch’s real-world performance is adherence, wear duration, early removal, replacement behaviour, application site and the incidence and severity of local skin reactions, and none of those appear in a controlled trial in the form they appear in use. Ask which of those are captured, by whom, at what interval and with what definition. A real-world dataset that records only prescription and refill has recorded dispensing rather than use, and the two differ most for exactly the products where adherence is the question.</p><p>Specification 4: The data quality standard is defined, including traceability and the ability to link. Real-world data are collected for care rather than for evidence, so the quality requirements have to be imposed rather than assumed: source documentation, an audit trail for any change, consistent coding of the product and the event, completeness checks and a mechanism for query resolution. Ask how the data are documented, whether the source can be verified, how the product is identified in each source, and whether records can be linked across sources without ambiguity. An unlinkable dataset cannot be reconciled, and an unreconcilable dataset cannot be defended.</p><p>Specification 5: The analysis plan is pre-specified, including the confounding control and the sensitivity analyses. Real-world comparisons are confounded by indication, by severity, by adherence and by the fact that the patients who persist with a patch are different from the patients who stop. Ask for the pre-specified estimand, the confounder list and how each was selected, the method used to control for confounding, and the sensitivity analyses that would test the conclusion. A plan written after the data are visible is a plan that has chosen its hypothesis from its results, which is the definition of a finding that will not replicate.</p><p>Specification 6: Regulatory acceptability is established per market for the specific purpose before the study starts. Regulators differ in what they will accept real-world evidence for, and the same dataset can be adequate for a safety update in one market and inadequate for a labelling change in another. Ask for the advice received or the precedent relied on in each market, for the specific purpose intended, and for the evidence standard that authority applies to that purpose. Ask also whether a scientific advice procedure is available and whether it was used, because the cheapest moment to discover that the design is unacceptable is before the first patient is enrolled.</p><p>Specification 7: The data ownership, the privacy basis and the buyer’s access are specified. Real-world data carry patient data, which means a legal basis for processing, a governance arrangement with each source, and rules on cross-border transfer where the sources and the analysis sit in different jurisdictions. Ask who owns the dataset, who may analyse it, who may submit it, what the buyer receives and when, and what happens to the dataset when the programme ends. A registry whose data belong to the site, or to the vendor, or to the manufacturer, is a registry whose evidence the buyer may be unable to use at the moment it is needed.</p><p>Real-world evidence is not a cheaper trial; it is a different instrument with a narrower range and a stricter requirement for pre-specification. Ask for the intended use and its limits, the sources and their bias, the patch-specific variables, the quality and linkage standard, the pre-specified analysis, the per-market acceptability and the ownership, and the registry becomes a study rather than a hope. The failure that costs the most is not a study that produced a negative result; it is a dataset collected for three years that cannot answer the question it was collected for, because nobody wrote the question down first.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a450abc1565c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Label Update Will Reach One Market Eleven Weeks Before Another: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-label-update-will-reach-one-market-eleven-weeks-before-another-seven-2da033e1a813?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/2da033e1a813</guid>
            <category><![CDATA[regulatory-affairs]]></category>
            <category><![CDATA[quality-management]]></category>
            <category><![CDATA[labeling-compliance]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[transdermal-patch]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Tue, 10 Nov 2026 01:52:44 GMT</pubDate>
            <atom:updated>2026-11-10T01:52:44.054Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Label Update Will Reach One Market Eleven Weeks Before Another: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>A label change on a transdermal patch looks like the smallest regulatory event a programme will ever manage, because the product does not change and the only thing that moves is the artwork. In practice the label is the part of the product that reaches the patient, the regulator and the border, and a change to it is a change to all three. The failure mode is rarely that the change is wrong; it is that the change is executed in one market and not another, or executed in the artwork and not on the shelf. The seven specifications below are what a buyer should require before contract signing, because a label update is a set of parallel executions rather than a single design task.</p><p>Specification 1: The reason for the change is classified, because the classification determines the process. A label change can originate in a new safety requirement, in a market-specific requirement such as a language or a pack configuration, or in a company decision such as a rebrand or a layout refresh, and the three have different pathways, different timelines and different implementation freedom. Ask for the reason, the evidence behind it, and which process the reason triggers. A programme that treats a safety-driven change with the same process as a layout refresh has conflated the two and will implement the safety change at the speed of the design cycle.</p><p>Specification 2: The change is governed as a change control event, not as a design event. A change control event has an impact assessment, an approval, a version history, an effective date and a closure with evidence, and it is the vehicle that connects the artwork file to the batch record and to the market. Ask whether the change is raised in the change control system, what the impact assessment covers, who approves it and how the effective date is communicated to manufacturing. A label change managed in a design tool is a change that will be discovered in a finished-goods warehouse.</p><p>Specification 3: The variation category is determined per market before the change is made. The same label change may be implementable immediately in one market, notifiable within a defined window in a second and subject to prior approval in a third, and the artwork should not be sent to print until that has been established. Ask for the category in each market, the implementation freedom in each, the fee and the review time, and the sequence in which the markets can be executed. A single print run prepared before the classification is a print run that may have to be destroyed.</p><p>Specification 4: A safety-driven change is implemented within the shortest applicable period in every market, not sequentially by convenience. Where the change adds or strengthens a warning, the markets that allow immediate implementation carry a continuing exposure for every day the old label remains in use, and the markets that require prior approval should be filed first rather than last. Ask what the shortest permissible date is in each market, how the markets are sequenced against it, and who owns the coordination. A safety update executed market-by-market in order of administrative convenience leaves the last market exposed for the whole duration of the exercise.</p><p>Specification 5: The artwork matrix is counted in markets multiplied by languages multiplied by pack configurations. One change to one sentence can generate a dozen artwork files, each with its own translation, its own regulatory wording, its own pre-press check and its own approval, and each of which can be independently wrong. Ask for the matrix, the owner of each cell, the review path and the mechanism that prevents a correction in one file from propagating to another or failing to propagate to the rest. The matrix is the part of a label change that determines the schedule, and it is the part most often estimated by counting labels instead of counting files.</p><p>Specification 6: The existing label inventory is dispositioned, including stock in transit, in the warehouse and on the shelf. Changing the artwork changes nothing about the product that is already printed, packed and shipped, and the old label will continue to reach patients until that inventory is consumed or withdrawn. Ask what the inventory position is at the effective date, whether the old stock may be sold through, what the regulatory position is on selling through in each market, and who decides to relabel, rework or destroy. A change implemented in the artwork and not in the inventory is a change that exists in the document system and not in the market.</p><p>Specification 7: The execution is evidenced per market, and the buyer can obtain the evidence. The proof that a label change has been implemented is not the approved artwork, it is the first commercial batch bearing the new label and the market’s acknowledgement where an acknowledgement is required. Ask how the first batch is identified, where the batch number is recorded against the change, how the acknowledgement is filed per market, and how the buyer can obtain that evidence for its own files. A change whose evidence exists only as an email chain is a change the buyer cannot demonstrate to anybody who asks.</p><p>A label change is the cheapest regulatory event to start and one of the more expensive to finish, because finishing it requires eight parallel executions in every market at once. Ask for the reason classification, the change control record, the variation category per market, the safety-change sequencing, the artwork matrix, the inventory disposition and the per-market evidence, and the change becomes a project with a schedule rather than a file that was updated. The risk that a label change carries is not that the new label is wrong. It is that the old label is still on a shelf somewhere, and nobody can say where.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2da033e1a813" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch PSUR Will Fail the Cumulative Test: Seven Specifications Every Buyer…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-psur-will-fail-the-cumulative-test-seven-specifications-every-buyer-73d189bfbfae?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/73d189bfbfae</guid>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[regulatory-affairs]]></category>
            <category><![CDATA[pharmacovigilance]]></category>
            <category><![CDATA[quality-management]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Mon, 09 Nov 2026 01:04:24 GMT</pubDate>
            <atom:updated>2026-11-09T01:04:24.676Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch PSUR Will Fail the Cumulative Test: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>The periodic safety update report is the document where a transdermal product’s cumulative safety experience has to be converted into a judgement, and it is usually produced by reformatting the last report and adding the new cases. That approach survives for as long as nothing has changed, and it fails the moment a regulator reads the benefit-risk section and asks what data the conclusion rests on. The seven specifications below are what a buyer should require before contract signing, because a periodic report that cannot demonstrate a cumulative assessment is a report that has documented activity rather than safety.</p><p>Specification 1: The reporting frequency is derived from each market’s requirements and the product’s status, not applied uniformly. Reporting intervals differ by jurisdiction and by product maturity: a newly authorised product is frequently reported more often than an established one, and the interval may be defined by the date of first authorisation in each market rather than by a single global calendar. Ask for the interval that applies in each market, the reference date from which it is counted, and what event would trigger a report outside the cycle, such as a new safety concern or a variation. A programme that runs one annual report for every market is a programme with a filing error waiting in at least one of them.</p><p>Specification 2: The report draws on the same dataset and the same case definitions as the surveillance system and the complaint system. A periodic report and a surveillance review that count adverse events differently produce two numbers about one product, and the reconciliation is normally performed after somebody notices. Ask what the case definition is, which system the report draws from, who reconciles the numbers when the two disagree, and whether the complaint data and the adverse event data are linked or parallel. Two parallel datasets for one product is a data integrity problem before it is a safety problem.</p><p>Specification 3: The report is cumulative in structure, and the interval data is presented inside the cumulative picture. A periodic safety report answers what the total experience looks like now, not what happened since the last report. Ask whether the report contains a cumulative table for the whole marketing life as well as an interval table, whether the cumulative exposure estimate is updated, and whether the reader can see the trend across periods rather than only the current slice. A report that presents interval data only has removed the denominator that makes a rate a rate.</p><p>Specification 4: The benefit-risk assessment states the method, not only the conclusion. Every periodic report concludes that the benefit still outweighs the risk, and the value of the report lies in what supports that conclusion. Ask what data were used, what the exposure denominator was and how it was estimated, what the assessed risks were, what measures were in place for each, and what would have changed the conclusion. A benefit-risk section that recites the indication and asserts the conclusion has performed the formatting of an assessment rather than the assessment.</p><p>Specification 5: Each signal carries its closure evidence, not a status label. A signal table that records closed against every line is a table that records nothing. Ask what each closed signal was assessed against, what investigation was performed, what the outcome was, whether the signal recurred after closure, and where the supporting record sits. Ask also whether signals closed in earlier periods are re-reviewed in the cumulative view, because a signal that recurs after closure is the strongest evidence that the original assessment was incomplete.</p><p>Specification 6: The report is consistent with the risk management plan and with the commitments made at registration. A periodic report and a risk management plan that describe different safety concerns for the same product cannot both be current, and the inconsistency is one of the simplest findings to make because both documents are in the file. Ask for the mapping between the report’s safety concerns and the risk management plan’s identified risks and missing information, and ask what happens when the report identifies a concern that the plan does not contain. The gap between the two is either an update that was not made or an assessment that was not done.</p><p>Specification 7: The submission responsibility, the signatory and the late-filing consequence are defined. A periodic report has a due date in each market, a responsible holder who submits it, and a signature that carries accountability for its content. Ask who prepares, who reviews, who signs and who submits in each market, what the escalation is when a deadline is at risk, and what the consequence of a late submission is in that jurisdiction. Ask also what the buyer can obtain afterwards, because the submitted report and its acknowledgement are the buyer’s evidence that the obligation was discharged.</p><p>A periodic safety report is an assessment with a formatting standard, and most programmes produce the formatting and skip the assessment. Ask for the interval per market and the trigger events, the single dataset, the cumulative structure, the benefit-risk method, the closure evidence, the consistency with the risk management plan and the submission accountability. A report built that way survives the question that a regulator actually asks, which is not whether the report was submitted but whether the product is still worth the risk it carries.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=73d189bfbfae" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Transdermal Patch Post-Marketing Surveillance Will Miss the Signal That Matters: Seven…]]></title>
            <link>https://medium.com/@kongdyhealth/why-your-transdermal-patch-post-marketing-surveillance-will-miss-the-signal-that-matters-seven-4d5465ab7be9?source=rss-d964862eb840------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d5465ab7be9</guid>
            <category><![CDATA[transdermal-patch]]></category>
            <category><![CDATA[regulatory-affairs]]></category>
            <category><![CDATA[pharmacovigilance]]></category>
            <category><![CDATA[contract-manufacturing]]></category>
            <category><![CDATA[quality-management]]></category>
            <dc:creator><![CDATA[KONGDY Health]]></dc:creator>
            <pubDate>Fri, 06 Nov 2026 00:50:44 GMT</pubDate>
            <atom:updated>2026-11-06T00:50:44.371Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Your Transdermal Patch Post-Marketing Surveillance Will Miss the Signal That Matters: Seven Specifications Every Buyer Should Require Before Contract Signing</h3><p>Post-marketing surveillance is the phase of a transdermal product’s life where the evidence stops being generated deliberately and starts arriving as a side effect of doing business, and almost every programme treats it accordingly. A plan is written for the regulatory submission, the plan is filed, and the operational reality becomes a complaint inbox and an annual report assembled in a hurry. The consequence is that the product is watched for the things the plan predicted and not for the things that actually happen, and the difference between those two sets is where a safety signal becomes a recall. The seven specifications below are what a buyer should require before contract signing, because a surveillance plan that is not executable is a document that describes vigilance without performing it.</p><p>Specification 1: The plan names the data sources, the frequency, the owner and the threshold, per market. A surveillance plan that says the product will be monitored is not a plan; a plan states which sources are read, how often each is read, who reads it, what constitutes a signal worth escalating and what happens when that threshold is crossed. Ask for the table: source, frequency, responsible person, escalation threshold, record of the review. Ask also whether the table differs by market, because a complaint channel that exists in one market does not exist in another and an assumption of symmetry is how a market goes unwatched.</p><p>Specification 2: Active and passive data sources are both identified, and the under-reporting gap is addressed. Passive surveillance is what comes in: complaints, spontaneous reports, returns, distributor feedback. Active surveillance is what is sought: structured follow-up with prescribers or pharmacies, user surveys, literature review, registry query. Ask which active sources are used, what proportion of the patient population they cover, and how the programme estimates the under-reporting rate. A programme that relies only on passive sources is measuring how often patients complain rather than how often the product fails.</p><p>Specification 3: The signal detection method is defined before the data arrive, including who looks and how often. Signal detection is a process rather than an intention: a defined dataset, a defined statistic or review method, a defined reviewer, a defined interval and a defined record of the conclusion. Ask what method is used, whether it is a simple review of case counts or a disproportionality calculation, what the review interval is, and where the conclusion of each review is documented. A programme with no written method detects signals when somebody happens to notice, which is the least reliable method available.</p><p>Specification 4: There is a defined escalation route from a signal into the deviation, complaint and CAPA systems, with a clock attached. A signal that is recognised and not escalated is a signal that has been wasted. Ask what happens when a signal is confirmed: which system receives it, who decides, how quickly, and what evidence closes it. Ask also whether the same signal arriving twice within a year is treated as a new event or as an escalation of the first, because programmes that treat each report as new repeatedly investigate the same failure and never fix it.</p><p>Specification 5: The surveillance performed matches the surveillance committed to in the registration. A product is approved with commitments: a specified post-market study, a specified registry, a specified periodic report, a specified follow-up period. Ask for the list of commitments made in each market and the evidence that each is being performed on schedule. The gap between a commitment and its performance is one of the easiest findings for an inspector to make, because the commitment is in the file and the performance is a date that has passed.</p><p>Specification 6: The surveillance data and the periodic report use the same source and the same definitions. A periodic safety report and an annual product review that count complaints differently produce two numbers that cannot both be right, and the reconciliation is usually performed after an inspector asks. Ask what the case definition is, what the counting rules are, whether the periodic report draws from the same dataset as the surveillance review, and who reconciles the numbers when they disagree. Two sets of numbers about the same product is a data integrity problem before it is a safety problem.</p><p>Specification 7: The data belongs to the buyer, can be produced on request, and survives the end of the manufacturing relationship. Surveillance data is the buyer’s evidence that it has discharged its obligations, and it is needed for the periodic report, for a regulatory inquiry, for a variation and for a due diligence. Ask who holds the raw case data, in what form, for how long, and how quickly it can be produced when a regulator asks; ask what happens to it when the manufacturing agreement ends and what the buyer may take away. A manufacturer that holds the only copy of the safety data has positioned itself as the buyer’s regulator, whether or not it intends to.</p><p>Surveillance is the cheapest form of risk management available to a patch programme, and it is the one most often reduced to a document. Ask for the source table, the active sources and the under-reporting estimate, the signal detection method, the escalation route with its clock, the commitment register, the single dataset and the data ownership, and the plan becomes an operation rather than a submission annex. The cost of an unexecuted plan is not paid at the inspection. It is paid at the point where a trend that would have been visible in a reviewed spreadsheet becomes visible in a cluster of patient reports, and the interval between those two moments is the whole value of surveillance.</p><p>For OEM/ODM transdermal patch manufacturing, visit <a href="http://www.kongdyhealth.com/">www.kongdyhealth.com</a> or contact kongdyhealth@gmail.com.</p><p>KONGDY Health has been an OEM/ODM manufacturer of transdermal patches since 1989.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d5465ab7be9" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>