In This Article
- Executive Summary
- Why Supervisors Decide Whether an eQMS Takes Hold
- Bringing Supervisors In Before Go-Live
- The Dashboards and Authority Supervisors Need
- Hypercare in the First 90 Days
- Measuring eQMS Adoption From the System Itself
- Retiring Paper Side-Systems and Old Habits
- When a Supervisor Is the Obstacle
- Conclusion
- For Further Reading
- References & Sources
Executive Summary
The day an electronic quality management system (eQMS) goes live is the day the project team starts to leave. From that point, eQMS adoption depends far more on front-line supervisors than on the vendor, the validation package, or the steering committee. Supervisors approve and return records, coach the people who create them, and decide, shift by shift, whether a paper log or a personal spreadsheet survives alongside the new system.
Most rollout plans treat supervisors as one more group of end users to train. This article argues for treating them as the owners of the change at the team level, and sets out what that takes in practice: involvement in design before go-live, dashboards and decision rights that let them act on what they see, a 90-day hypercare period built around a daily supervisor huddle, adoption measures drawn from the system’s own records, and a plan for retiring paper side-systems rather than hoping they fade.
The sections that follow cover why supervisors matter under GMP as well as for change, how to prepare them, what they need on day one, how to run the first 90 days, which numbers show whether the system is in use, how to deal with old habits, and what to do when a supervisor is the source of resistance.
Why Supervisors Decide Whether an eQMS Takes Hold
An eQMS replaces paper and spreadsheet processes for deviations, CAPA (corrective and preventive action), change control, document control, complaints, audits, and training records. Selection and validation take months. The go-live date gets a great deal of attention. Then the system is handed to the people who will use it every day, and the questions that decide whether the program succeeded start to be answered on the floor, in the lab, and in the quality office, one record at a time.
The person who answers most of those questions is the front-line supervisor: the production shift lead, the QC laboratory supervisor, the warehouse lead, the QA specialist who leads a small review team. They are rarely on the steering committee. They are often trained last. Yet nearly every adoption outcome passes through them.
The Three Things a Supervisor Does Every Shift
Supervisors influence adoption through three ordinary parts of their job.
- They approve and return records. When a supervisor reviews a deviation initiation, a logbook entry, or a training assignment in the new system, that review sets the real standard. If incomplete entries are approved, the team learns that the fields do not matter. If records are returned with clear reasons, the team learns what good looks like.
- They coach. Most questions about a new system are asked of the nearest experienced person, not the help desk. A supervisor who can answer “where do I attach the chromatogram” in thirty seconds keeps the work moving inside the system. A supervisor who says “just write it down and we will enter it later” starts a side-system.
- They decide whether workarounds survive. Every new system produces gaps: a field that does not fit a real situation, a step that takes too long, a report that does not exist yet. Someone on the team will find a way around it. The supervisor decides, often without saying anything, whether that workaround is tolerated, escalated, or stopped.
None of these is a project task. All of them happen after the project team has moved on. That is why a rollout can pass every validation test and still fail to change how work is done.
What the Regulations Already Assume About Supervisors
The supervisor’s role is not only a matter of change management. GMP rules already give supervisors a formal place in the record. Under 21 CFR 211.188, batch production and control records must include identification of the persons performing and directly supervising or checking each significant step in the operation.3 FDA’s data integrity guidance makes the same point for record review. It says that “the regulations provide flexibility to have some activities reviewed by a person directly supervising or checking information,” citing 211.188, and that personnel responsible for record review should review the audit trails that capture changes to data as they review the rest of the record.4 The guidance also offers a plain comparison: “Audit trail review is similar to assessing cross-outs on paper when reviewing data.”4
For an eQMS, this matters in a specific way. The supervisor who used to review a paper form for cross-outs and late entries now has to review the electronic record and its audit trail. That is a new skill, and it is one the regulations expect them to perform well. A supervisor who approves records without looking at the history is not only slowing adoption. They are weakening a control that inspectors will test.
ICH Q10, the harmonized guideline on pharmaceutical quality systems, places responsibility for defining roles, responsibilities, and authorities with management, and asks management to make sure these are communicated and understood at all levels of the company.5 It also asks management to show strong and visible support for the quality system and to make sure it is implemented throughout their part of the organization.5 Supervisors are the level of management closest to the work. When senior leaders sponsor an eQMS and supervisors do not visibly use it, staff draw their conclusions from the supervisor.
What the Change Research Says
Change management research points the same way. Prosci’s benchmarking work has named engagement with people managers a top contributor to change success since 2011, and its research shows that a majority of employees prefer to receive personal messages about changes directly from their managers.1 Prosci describes five roles for people managers during a change: communicator, liaison, advocate, resistance manager, and coach.1 In its Best Practices in Change Management 12th Edition study, 63 percent of respondents said their organization does not adequately prepare people managers with the skills, training, and tools they need to lead during change.1 Prosci adds that coaching and managing resistance are the roles that depart most from what managers have historically done.1
Gallup’s research on managers gives a broader view of how much the immediate manager shapes a team. Gallup estimates that managers account for at least 70 percent of the variance in employee engagement scores across business units.2 Engagement and system adoption are different things, and that figure should not be read as a measure of eQMS success. It does show that the person a team reports to has an outsized effect on how that team behaves, which is the premise of this article.
The Supervisor Sets the Practice, for Better or Worse
FDA enforcement records show how much a front-line lead can shape what a team does with its records, in either direction. In a June 2026 warning letter to a drug manufacturer in South Korea, FDA described a microbiology team leader who had instructed staff to discard bioburden plates with significant growth, and who later admitted to removing completed logbook pages with a knife. The same letter reports that investigators found 1,897 blank uncontrolled CGMP-related forms during a walkthrough of the microbiology laboratories.6 The firm’s response, as FDA summarized it, included the departure of several people involved and the creation of a data integrity team to oversee microbiology laboratory data practices.6
That case involved misconduct, not a struggling system rollout, and it would be wrong to draw a straight line from it to eQMS adoption. The point is narrower. What a team treats as “the record” is set by the person who leads it. An eQMS can make good practice easier, but it cannot make a supervisor use it. Rollout plans need to treat supervisors as the people who will either carry the new way of working into daily practice or keep the old one alive.
Bringing Supervisors In Before Go-Live
The most common mistake is timing. Supervisors are often invited to a demonstration near the end of configuration, trained a week or two before go-live alongside their teams, and then expected to lead adoption from day one. By then the design decisions that will shape their daily work have been made without them. They learn the system at the same time as the people who will come to them for help.
Start With Their Records, Not the Vendor Demo
Supervisors know where the current process breaks. They know which deviation categories are always argued over, which CAPA actions never close on time, and which training assignments get missed when someone changes shifts. That knowledge is most useful during configuration, not after it.
A practical way to get it is to run design walkthroughs using a supervisor’s own recent records. Take five real deviations from the last quarter, with names removed if needed, and walk them through the configured workflow. Ask the supervisor where they would expect to find each piece of information, what they would check before approving, and where the new workflow adds steps they do not see the point of. This exposes gaps that a requirements workshop rarely finds: a mandatory field that does not apply to half the records, a routing rule that sends every minor deviation to the same QA reviewer, a status name that means something different on the shop floor.
Prosci makes a related point about the liaison role: people managers are in the best position to give the project team design input, usability results, and employee comments, and to identify functionality needs and concerns during implementation.1 A supervisor who saw their own feedback change the configuration has a reason to defend the system later. A supervisor who was shown a finished product does not.
Make Them the First Trained, Not the Last
Supervisors should be trained before their teams, and trained differently. End-user training teaches people how to create and complete records. Supervisor training needs to cover what end users learn plus four more things:
- How to review a record and its audit trail. What changed, when, and by whom, and what a pattern of late or repeated edits looks like in the new system.
- How to return a record well. The difference between a return that teaches (“root cause section describes the event, not the cause; see SOP section 6.2”) and one that frustrates (“incomplete”).
- How to use the team-level views. Queues, aging reports, and overdue items for their own team, not the site.
- How to escalate a system problem. Who to tell, what information to include, and how quickly to expect an answer.
Training supervisors two to three weeks early also gives them time to use the system in a training environment with realistic data. Supervisors who have closed ten practice records will answer their team’s questions with confidence. Supervisors who have watched a slide deck will not.
Choose Super Users Who Want the Role
Many programs appoint “super users,” team members with extra training who support their peers during go-live. Supervisors usually pick them, and the way they pick them matters. A study of electronic health record implementation by Yuan, Bradley, and Nembhard compared two hospital units. On one unit, super users had volunteered; on the other, managers had appointed them for their technology skills. The volunteer super users were more proactive, explained the reasons behind system steps rather than only the steps, and shared information with everyone on the unit, and clinicians on that unit showed greater improvement in information systems proficiency.7 The authors concluded that managers should allow super users to volunteer, because “the choice to participate is a greater predictor of super users’ effectiveness than their technological savvy.”7
The same study found that super users on both units also behaved in ways that hindered implementation, including creating workarounds.7 That finding is worth repeating to every supervisor before go-live. The most helpful person on the team can also be the one who invents the shortcut that everyone copies. Super users need a clear channel to report problems and a clear rule that workarounds are reported, not taught.
The study was small (two units in one hospital, 43 clinicians surveyed) and its setting is clinical rather than GMP. It is still one of the few published comparisons of how super users are selected, and its conclusion matches what most quality leaders see in practice.
A Supervisor Readiness Check
Before go-live, each supervisor should be able to answer yes to a short set of questions. Where the answer is no, the gap is a go-live risk for that team, and it should be tracked like any other.
| Readiness Question | Why It Matters | If the Answer Is No |
|---|---|---|
| Have I completed practice records in the training environment for every record type my team creates? | Supervisors answer most first-week questions. | Schedule hands-on time before go-live; do not rely on classroom training. |
| Do I know how to read the audit trail for a record I approve? | Record review under GMP includes the audit trail. | Add a short audit trail review exercise to supervisor training. |
| Can I see my team’s open, overdue, and returned records in one view? | Without a team view, supervisors manage from memory or a spreadsheet. | Configure the view before go-live, not as a later enhancement. |
| Do I know who decides on a configuration problem, and how fast? | Slow answers create workarounds. | Publish the hypercare triage contact and response targets. |
| Do I know which paper forms and logs are retired on go-live day? | Unclear retirement keeps paper alive. | Issue a retirement list per team, signed off by the supervisor. |
| Has my workload been adjusted for the first month? | Coaching and reviewing take time the normal schedule does not include. | Agree temporary cover or reduced targets with the area manager. |
The last question is the one most often skipped. A supervisor who is expected to hit normal production or testing targets while also coaching a team through a new system will choose the targets. That is a reasonable choice for them to make, and it is a planning failure, not a supervisor failure.
The Dashboards and Authority Supervisors Need
Many eQMS dashboards are built for site quality heads and steering committees. They show site-wide counts of open deviations, overdue CAPAs, and change controls by status. These are useful views, but they do not help a supervisor run a team on a Tuesday morning. A supervisor needs to see their own team’s work, and they need the authority to act on what they see.
The Views That Matter on Day One
Four views cover most of what a supervisor needs in the first months.
My Team’s Open Records
Every open record created by or assigned to the team, with current step, owner, and days at that step. Sorted by age at step, not by creation date, so that stuck records rise to the top.
Records Waiting on Me
Approvals and reviews assigned to the supervisor personally, with due dates. If this list grows, the supervisor is the bottleneck, and they should see that before anyone else does.
Returned and Reworked Records
Records sent back at any review step in the last 30 days, with the reason. This is the supervisor’s best coaching list: it shows which people and which record sections need help.
Training Due for My Team
Open training assignments, especially those tied to new or revised procedures. In many eQMS programs, training records are among the first modules live, and overdue training is an inspection finding of its own.
These views should be configured and tested before go-live, with real supervisors checking that they show the right people and records. A view that shows the wrong team, or that a supervisor cannot filter to their shift, will be abandoned in favor of a spreadsheet within a week.
Authority to Act, Not Just to See
A dashboard that shows a problem a supervisor cannot fix produces frustration, not adoption. Supervisors need defined authority in four areas.
- Reassigning work within the team. When someone is on leave or overloaded, the supervisor should be able to reassign a task without raising a ticket with the system administrator.
- Delegating approvals during absence. Most eQMS platforms support delegation. The rules for who may delegate to whom should be set in procedure and configuration before go-live, because approval delays during holidays are one of the most common early complaints.
- Raising configuration issues through a fast lane. A supervisor who reports that a field does not fit a real situation should get a decision within days during hypercare. The decision may be “use field X for now; we will change it in the next release,” but it needs to be a decision.
- Stopping a workaround. Supervisors should know they are expected to stop a workaround they see, and that they will be backed when they do.
A Decision Rights Table
Writing decision rights down avoids the slow, uncertain escalation that sends people back to paper. A simple table, agreed between quality, the business areas, and the system owner, is enough.
| Decision | Supervisor | System Owner or Business Process Owner | Quality Unit |
|---|---|---|---|
| Reassign a task within the team | Decides | Informed through audit trail | None |
| Delegate own approvals during absence | Decides within procedure limits | Maintains delegation rules | Approves the rules |
| Return a record for rework | Decides | Monitors return rates | Reviews trends |
| Interim guidance on using a field or step | Requests | Decides during hypercare | Approves if it affects GMP content |
| Configuration change to workflow | Requests and supplies examples | Assesses and plans | Approves under change control |
| Use of a paper downtime form | Starts downtime procedure | Confirms system outage | Reviews reconciliation |
| Retire a local paper form or log | Confirms team readiness | Confirms system covers the need | Approves retirement |
EU GMP Annex 11, the EU guidance on computerized systems, calls for close cooperation among relevant personnel such as the process owner, system owner, Qualified Persons, and IT, and for all personnel to have defined responsibilities and appropriate access for their duties.9 A decision rights table is one simple way to show that those responsibilities have been defined for the people who use the system most.
Hypercare in the First 90 Days
Hypercare is the period of extra support right after go-live, when the project team, vendor, and super users stay close to users, fix problems fast, and watch for signs that the system is not being used as designed. In many eQMS programs, hypercare is defined as two to four weeks of help desk coverage. That is usually too short and too narrow. The first records in a new eQMS often take weeks to move through every step, so many problems only appear once records reach approval, effectiveness checks, or closure.
Published evidence from regulated pharma rollouts is thin, but health record implementations offer a useful comparison. In a large academic medical center’s electronic medical record rollout, Rizer and colleagues reported that with sufficient support at go-live, intense support could often be limited to four to six weeks, depending on how often users worked in the system, and that support should then be tapered gradually so users did not feel abandoned.8 They also found that returning to clinics every two to three months during the first year worked well.8 An eQMS has a different user population and slower record cycles, but the pattern holds: intense support early, a planned taper, and scheduled return visits.
Three Phases Across 90 Days
Days 1 to 14: Presence
Super users and project staff are physically present in each area during working hours, including off shifts. Supervisors hold a short daily huddle with their team. The system owner holds a daily triage call with supervisors and the vendor. The goal is that no one leaves a record half-done or goes back to paper because they could not get an answer.
Days 15 to 45: Stabilization
In-person support tapers to scheduled hours and on-call. Triage moves from daily to three times a week. Supervisors start using the returned-records view to coach. The first records reach later workflow steps (QA approval, CAPA effectiveness checks), and new problems appear there. Interim guidance issued in phase one is reviewed: confirmed, replaced with a configuration change, or withdrawn.
Days 46 to 90: Ownership
Supervisors run adoption for their teams with support on request. Weekly adoption measures (see the next section) are reviewed team by team. Remaining paper side-systems are identified and given retirement dates. The program prepares an exit review that decides whether hypercare ends, continues for specific areas, or hands open items to the normal change control process.
Sorting Issues Into Four Bins
A large share of hypercare effort goes into deciding what kind of problem a report describes. Every issue should be sorted into one of four bins, because each needs a different owner and a different fix.
| Bin | Example | Owner | Typical Fix |
|---|---|---|---|
| System defect | Approval button errors for one user role | Vendor and system owner | Fix under change control, with impact assessment |
| Configuration gap | Deviation category list missing a common event type | System owner and business process owner | Interim guidance now, configuration change later |
| Training gap | Users attaching evidence in the wrong section | Supervisor, with super user help | Targeted coaching; update job aid |
| Procedure mismatch | SOP still describes paper routing | Procedure owner and quality unit | Revise SOP; interim memo if needed |
The procedure mismatch bin is easy to miss. If the SOP still describes the paper process, a careful employee is right to follow it, and a supervisor cannot fairly correct them. Procedure updates should be planned to take effect on go-live day, and any found later should be treated as urgent.
The Daily Supervisor Huddle
The most useful hypercare meeting is a short daily call between the system owner and the supervisors of every area that went live. Fifteen minutes is enough if the format is fixed:
- Each supervisor names any record stuck for more than two days and why.
- Each supervisor names any workaround they have seen or heard about.
- The system owner reports decisions on issues raised earlier, in the four bins above.
- One piece of interim guidance, if any, is agreed and sent in writing to all supervisors the same day.
This does two things. It gets problems to the people who can fix them quickly, which removes the main reason for workarounds. It also tells supervisors, every day, that reporting a workaround is expected and useful, not a confession of failure.
Write the interim guidance down. In the first weeks, many problems are solved by telling people to use a field or step in a particular way until a configuration change is made. That guidance changes how GMP records are completed. It should be written, approved by quality where it affects GMP content, dated, and withdrawn on a set date. Verbal guidance passed from supervisor to supervisor becomes a local rule that no one can trace.
Ending Hypercare on Evidence, Not the Calendar
Hypercare should end when the evidence says the system is stable and in use, not on a date set in the project plan. ICH Q10 says that after implementation, “an evaluation of the change should be undertaken to confirm the change objectives were achieved and that there was no deleterious impact on product quality.”5 An eQMS go-live is a change to the quality system itself, so this evaluation applies. Annex 11 also calls for computerized systems to be evaluated periodically, including deviation records, incidents, and problems, to confirm they remain valid and GMP compliant.9 The hypercare exit review is a natural first input to both.
A reasonable exit review asks four questions for each area: Are records moving through every workflow step without help? Are adoption measures stable or improving? Are all known side-systems retired or scheduled for retirement? Are open issues assigned to normal change control with owners and dates? An area that cannot answer yes to all four stays in hypercare, even if others exit.
Measuring eQMS Adoption From the System Itself
Login counts and training completions show that people can get into the system. They do not show that the system has replaced the old way of working. An earlier Sakara Digital article on behavioral KPIs set out general measures for any system rollout, including rework rate, workaround rate, and residual work in the old system. This section applies that thinking specifically to an eQMS, and adds one change: every measure is reported by team and by supervisor, because that is where it can be acted on.
The best source for these measures is the eQMS itself. Every record has timestamps for each step, every return has a reason, and every change has an audit trail entry. That data exists from day one, and using it avoids asking already busy supervisors to fill in surveys.
Cycle Time by Stage, Not Overall
Overall cycle time (from deviation initiation to closure, for example) is a lagging measure. In the first 90 days most records are not closed yet, so the overall number says little. Cycle time by stage is more useful: time from event to record initiation, time from initiation to first supervisor review, time in QA review, time from investigation completion to approval.
Two stages deserve close attention from supervisors.
- Event to initiation. If the time between when an event happened and when the record was opened is growing, people are holding information somewhere else first: a notebook, an email, a whiteboard. This is often the first sign of a paper side-system.
- Waiting for supervisor review. If records wait longest at the supervisor’s own step, the supervisor is the bottleneck. This is fixable (delegation, workload relief, a daily review slot), but only if the supervisor sees it.
Rework and Return Rates
Every time a reviewer sends a record back, the eQMS records it. The return rate (returned records as a share of submitted records, by step and by team) is one of the most direct adoption measures available. It should be expected to rise in the first weeks as reviewers learn what the new system shows them, then fall as teams learn what reviewers expect.
Two patterns should prompt a closer look. A team whose return rate never falls may need more coaching, or may be facing a configuration problem that makes correct completion hard. A team whose return rate is near zero from day one may have a reviewer who is approving without reading. The second pattern is the more serious one under GMP, and it can only be seen if return rates are reported by reviewer as well as by team.
Workaround Signals in the Data
Workarounds leave traces in eQMS data, even when no one reports them. Common signals include:
- Batch entry. Many records created by one person in a short window at the end of a shift, which suggests information was gathered elsewhere and entered later.
- Free-text overuse. The “other” category or a general comments field used far more than expected, which often means the configured options do not fit real events.
- Attachments that replace fields. A scanned paper form attached to a record in place of completed structured fields.
- Repeated edits after approval requests. Records changed many times between submission and approval, visible in the audit trail.
- Unusual approval timing. Large numbers of approvals in a few minutes, which may mean records are not being reviewed.
None of these proves a problem. Each is a reason for the supervisor, or the system owner, to look at a sample of records and ask the team what is happening. Alter’s theory of workarounds, published in Communications of the Association for Information Systems, is useful background here: it treats workarounds as the result of identifiable factors that determine whether a workaround is considered and how it is carried out, and it argues that they can be classified and analyzed rather than simply forbidden.10 A workaround signal is often the first evidence of a configuration gap.
Do not turn these signals into targets. If supervisors are scored on low return rates, return rates will fall, and review quality will fall with them. If teams are scored on short event-to-initiation times, records will be opened early with little in them. Use the measures to decide where to look, not to rank people.
Reading the Numbers by Team
A simple weekly view, one row per team, gives supervisors and the system owner a shared picture. The table below shows the measures and how to read them during the 90 days. It contains no benchmark values, because the right range depends on the record type, the site, and the previous process. Each site should set its own baseline from the first two weeks.
| Measure | Source in the eQMS | Expected Shape in 90 Days | What a Bad Pattern Suggests |
|---|---|---|---|
| Event to initiation time | Event date field and record creation timestamp | Stable or falling | Information held elsewhere first |
| Time waiting at supervisor review | Workflow step timestamps | Falling after week two | Supervisor overloaded or avoiding the system |
| Return rate at first review | Workflow return actions and reasons | Rises, then falls | Never falls: training or configuration gap. Near zero: review not happening |
| Share of records using “other” or free text | Category and comment fields | Falling as categories are refined | Configured options do not fit real events |
| Records with scanned paper attachments in place of fields | Attachment metadata and sampling | Near zero after first month | Paper form still in use |
| Help requests per user | Help desk tickets by area | Falling, with themes changing | Same question repeated: job aid or design issue |
The point of reporting by team is to give each supervisor their own numbers. A site-wide average hides the one team that is still running on paper. A team row lets the supervisor see it, and lets the system owner offer help to that team specifically.
Retiring Paper Side-Systems and Old Habits
A paper side-system is any record kept outside the eQMS that holds information the eQMS is supposed to hold: a deviation notebook kept at the line, a spreadsheet tracking CAPA due dates, a whiteboard of open investigations, a binder of printed training matrices. Most are created with good intentions. Some predate the eQMS and were never retired. Others appear after go-live because the new system did not meet a need fast enough.
Why Side-Systems Appear
Research on workarounds in other regulated, high-pressure settings helps explain how they start and spread. A scoping review of nurses’ workarounds in acute care by Debono and colleagues, covering 58 studies, found that workarounds are shared or passed on informally, particularly from senior to junior staff, and become part of the group behavior.11 The review linked their development to factors including heavy workloads, productivity pressures, and a negative organizational climate characterized by poor leadership and a lack of staff involvement in decision-making.11 Its overall conclusion was that “workarounds enable, yet potentially compromise, the execution of patient care.”11
The same pattern shows up after an eQMS go-live. A side-system usually starts because a person needs something the system does not yet give them: a quick view of what is open, a place to note an event before they have time to enter it, a way to track something the configuration missed. It spreads because it works and because a respected colleague uses it. It survives because no one with authority tells the team to stop, or because no one offers a better option. Supervisors are the people best placed to see all three stages.
What Regulators Say About Uncontrolled Records
A side-system that holds GMP information is not only an adoption problem. It is a data integrity risk.
FDA’s data integrity guidance says that blank forms, including electronic worksheets and laboratory notebooks, “should be controlled by the quality unit or by another document control method,” and it gives the example of bound, paginated, stamped notebooks because they “allow easy detection of unofficial notebooks as well as any gaps in notebook pages.”4 It also states that all data required to recreate a CGMP activity should be maintained as part of the complete record.4
The PIC/S guidance on data management and integrity (PI 041-1) is direct about hybrid systems, meaning combinations of paper and electronic records: “the use of hybrid systems is discouraged and such systems should be replaced whenever possible.”12 Where hybrid arrangements remain, PIC/S expects procedures and records that control the interface between manual and automated systems, including manual input of manually generated data into computerized systems.12 PI 041-1 is guidance, not law (it says so itself), but it reflects how PIC/S member inspectorates approach the topic.12
Recent enforcement shows what happens when unofficial records hold real data. In a March 2026 warning letter to a contract testing laboratory in India, FDA reported that “laboratory staff used unofficial personal diaries to record procedures, analytical observations, results, method modifications, and deviation descriptions,” and that the firm “lacked procedures to control the use of these personal diaries and to ensure proper retention of all CGMP data.”13 That letter describes far more than a rollout problem, including discarded analytical records. But the diaries finding is the side-system risk in its plainest form: deviation descriptions and results recorded in a place the quality system could not see.
A Downtime Procedure Is Not a Side-System
Not all paper is a problem. Annex 11 expects firms to make provisions for continuity of critical processes if a system breaks down, for example a manual or alternative system, and says these arrangements should be documented and tested.9 A controlled downtime form, issued under a procedure, reconciled when the system is back, and entered into the eQMS with a clear link to the paper original, is a legitimate part of the design.
The difference between a downtime procedure and a side-system is control. The table below gives supervisors a simple test.
| Question | Controlled Downtime Record | Side-System |
|---|---|---|
| Is it described in an approved procedure? | Yes | No, or only in local habit |
| Is it used only when the system is unavailable? | Yes, with start and end recorded | Used every day, or whenever convenient |
| Are the forms issued and reconciled? | Yes, numbered and reconciled | Photocopied or printed as needed |
| Does every entry reach the eQMS? | Yes, with a link to the original | Some do, some do not, no one checks |
| Does quality review it? | Yes, as part of reconciliation | No |
Retiring a Side-System Without a Fight
Banning side-systems by memo rarely works. People keep them because they meet a need, and a ban without a replacement moves them out of sight. A better approach treats each side-system as a requirement that has not yet been met.
Find Them by Asking, Not Auditing
In the second month of hypercare, ask each supervisor to walk their area with a super user and list every place information is recorded outside the eQMS. Make clear that finding a side-system is useful, not a failure. Teams hide what they expect to be punished for.
Understand the Need Each One Meets
For each side-system, ask what it does that the eQMS does not. Often the answer is a view (a list of what is open for my shift), a timing problem (I cannot get to a terminal until break), or a missing field. Each of these points to a fix.
Fix the Gap or Formalize the Practice
If the eQMS can meet the need through a report, a view, a mobile device, or a configuration change, plan that work. If the need is legitimate and cannot be met yet, bring the practice under control: an issued, reconciled form with a procedure, reviewed by quality, until the system catches up.
Set a Retirement Date and Have the Supervisor Remove It
Every side-system gets a retirement date tied to its fix. On that date, the supervisor removes it from the area, in front of the team, and any information still in it is transferred and reconciled. The supervisor doing this, not a quality auditor, is what tells the team the change is real.
Old habits that are not physical records need the same treatment. A supervisor who still asks for a verbal update before opening the system, or who prints records to review them on paper, is teaching the team that the eQMS is secondary. These habits are best raised privately, and early, by the supervisor’s own manager.
When a Supervisor Is the Obstacle
Most supervisors want the new system to work, even when they are frustrated with it. Some do not, and a few will slow adoption for their team without ever saying they object. Leaders need a way to tell the difference between a supervisor who is overloaded, one who has a fair complaint about the system, and one who is resisting.
Signs of Passive Resistance
Passive resistance rarely looks like refusal. It looks like a team whose records always wait longest at the supervisor step, a supervisor who never raises issues in the daily huddle while their team’s help requests are the highest on site, a paper log that keeps reappearing after it was retired, or a supervisor who describes the system to their team in terms that make it sound temporary (“for now we have to put it in the system”). The adoption measures in the previous section will often show the pattern before anyone names it.
Separating Workload From Resistance
Before treating a supervisor as resistant, check the simpler explanations. Is their team the largest, or the one with the most record types? Did their workload relief for the first month happen? Are their complaints about the system accurate? A supervisor who says the deviation form does not fit their area’s events may be right, and the free-text rate for their team will show it.
Prosci’s research identifies coaching and resistance management as the manager roles that depart most from what managers have historically done, and notes that both require new skills.1 That applies to the managers of supervisors too. An area manager who does not know how to coach a supervisor through a change will tend either to ignore the problem or to escalate it too quickly. Both make it worse.
A Short Escalation Path
A simple path, agreed in advance, helps.
- Data first. The system owner shares the team’s adoption measures with the supervisor privately, and asks what is behind them.
- Remove the obstacles that are real. Workload relief, a configuration fix, extra super user time, or a training session for the team.
- Agree specific expectations. For example: review queue cleared daily, side-system retired by a set date, issues raised in the huddle.
- Involve the supervisor’s manager. If the pattern continues after obstacles are removed, it becomes a performance conversation owned by the line manager, not the project team.
Where the pattern involves approving records without review or keeping GMP data outside the controlled system, it is no longer only an adoption matter. It should be handled through the quality system like any other compliance concern.
What Senior Leaders Owe Supervisors
Asking supervisors to own adoption is a real request, and it comes with obligations. Senior leaders need to give them time (workload relief in the first month), tools (team-level views that work on day one), authority (a written decision rights table), and fast answers (a hypercare triage that decides within days). They also need to back supervisors who stop a workaround or retire a side-system, especially when the person using it is senior or well liked.
ICH Q10 asks management to make sure there is a timely and effective process for communicating and escalating quality issues to the appropriate levels.5 For an eQMS rollout, the most important direction of that flow is upward from supervisors. If the daily huddle shows supervisors raising problems and getting decisions, the system will be adopted. If supervisors learn that raising problems produces nothing, they will solve the problems themselves, usually with paper.
A simple test of supervisor ownership. Ninety days after go-live, ask each supervisor three questions: Where do you look to see what your team has open? What did you change in the system, or ask to change, in the last month? What paper did you retire? A supervisor who answers “the eQMS dashboard,” gives a specific example, and names a retired form owns the change. A supervisor who cannot answer is still waiting for someone else to own it.
Conclusion
eQMS adoption is decided after go-live, in thousands of small choices about how records are completed, reviewed, and returned, and about whether a paper log stays on the bench. Supervisors make or shape most of those choices. GMP rules already give them a formal place in the record as the people directly supervising or checking work, and FDA expects record reviewers to read the audit trail. A rollout that trains supervisors last, gives them site-level dashboards they cannot act on, and ends hypercare after two weeks leaves them to manage the hardest part of the change with the least support.
The alternative is not complicated: bring supervisors into design with their own records, train them first and differently, choose super users who want the role, give supervisors team-level views and written decision rights, run 90 days of hypercare around a daily supervisor huddle, measure adoption from the system’s own data by team, and retire side-systems by meeting the need behind them. Sakara Digital works with pharma and biotech organizations planning and stabilizing quality system rollouts. If you are preparing for an eQMS go-live, or working through the months after one, and want an independent view on where adoption is at risk, we are happy to have that conversation.
For Further Reading
For Further Reading
- Behavioral KPIs: Why Your Digital Transformation Metrics Are Wrong
- Why AI Adoption in Quality Stalls After the Pilot
- Cleaning Legacy Deviation and CAPA Data Before an eQMS Migration
- Choosing Between an eQMS Module and a Best-of-Breed Tool
- Training Effectiveness Beyond the Read-and-Understand Signature
- Change Management for Digital Transformation in Pharma
References & Sources
- Kempton, Lisa. “CLARC: The Role of People Managers in Change Management.” Prosci, updated August 7, 2026. https://prosci.com/blog/enabling-managers-to-become-great-change-leaders
- Beck, Randall, and Jim Harter. “Managers Account for 70% of Variance in Employee Engagement.” Gallup Business Journal, April 21, 2015. https://news.gallup.com/businessjournal/182792/managers-account-variance-employee-engagement.aspx
- U.S. Code of Federal Regulations. 21 CFR 211.188, Batch Production and Control Records. eCFR. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-211/subpart-J/section-211.188
- U.S. Food and Drug Administration. “Data Integrity and Compliance With Drug CGMP: Questions and Answers.” Guidance for Industry, December 2018. https://www.fda.gov/media/119267/download
- International Council for Harmonisation. “ICH Q10: Pharmaceutical Quality System.” Step 4, June 2008. https://database.ich.org/sites/default/files/Q10%20Guideline.pdf
- U.S. Food and Drug Administration. Warning Letter 320-26-95 to Huons Co., Ltd., June 15, 2026. https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/huons-co-ltd-724650-06152026
- Yuan, Christina T., Elizabeth H. Bradley, and Ingrid M. Nembhard. “A Mixed Methods Study of How Clinician ‘Super Users’ Influence Others During the Implementation of Electronic Health Records.” BMC Medical Informatics and Decision Making 15, no. 26 (2015). https://pmc.ncbi.nlm.nih.gov/articles/PMC4407776/
- Rizer, Milisa K., Beth Kaufman, Cynthia J. Sieck, Jennifer L. Hefner, and Ann Scheck McAlearney. “Top 10 Lessons Learned from Electronic Medical Record Implementation in a Large Academic Medical Center.” Perspectives in Health Information Management (2015). https://pmc.ncbi.nlm.nih.gov/articles/PMC4558484/
- European Commission. “EudraLex Volume 4, Annex 11: Computerised Systems.” January 2011. https://health.ec.europa.eu/system/files/2016-11/annex11_01-2011_en_0.pdf
- Alter, Steven. “Theory of Workarounds.” Communications of the Association for Information Systems 34 (2014). https://aisel.aisnet.org/cais/vol34/iss1/55/
- Debono, Deborah S., David Greenfield, Joanne F. Travaglia, Janet C. Long, Deborah Black, Julie Johnson, and Jeffrey Braithwaite. “Nurses’ Workarounds in Acute Healthcare Settings: A Scoping Review.” BMC Health Services Research 13, no. 175 (2013). https://pmc.ncbi.nlm.nih.gov/articles/PMC3663687/
- Pharmaceutical Inspection Co-operation Scheme. “PI 041-1: Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments.” July 1, 2021. https://picscheme.org/docview/4234
- U.S. Food and Drug Administration. Warning Letter 320-26-47 to Tentamus India Private Limited, March 3, 2026. https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/tentamus-india-private-limited-720463-03032026








Your perspective matters—join the conversation.