Biology, chemistry, the lab mindset, the regulated routes in, and the rules that govern real labs.

Foundations for Life Science & Biotech Careers

## Why it matters

Lab job postings are full of acronyms: "GMP experience preferred", "CLIA-certified laboratory", "GLP study support". They name the rule sets that decide how a lab must work, how it is inspected and what its records must show. Knowing which rules apply to which lab tells you what a job will be like and what training you will need. This lesson covers the US framework at awareness level; the regulations themselves are long, change over time, and are taught in detail by employers.

## GLP: Good Laboratory Practice

**What it covers:** non-clinical laboratory studies (often animal or lab studies of toxicity and safety) whose results are submitted to regulators to support approval of products such as drugs, medical devices, food additives and pesticides.

**Who sets it:** in the US, the FDA's GLP regulations are in **21 CFR Part 58**, and the EPA has its own GLP rules for pesticides and chemicals. Internationally, the **OECD** Principles of GLP play the same role.

**What it requires, in outline:**

- A named **study director** responsible for each study.
- An independent **quality assurance unit** that audits studies.
- Written **protocols** and **SOPs**, followed and documented.
- Calibrated, maintained equipment and characterised test materials.
- **Archived raw data** and specimens so a study can be reconstructed years later.

GLP is about the **integrity of a study**. It does not tell scientists which experiments to run.

## GMP: Good Manufacturing Practice

**What it covers:** the manufacture, testing and release of products such as drugs, biologics and medical devices, so every batch is safe and consistent. You will often see "cGMP" (current GMP).

**Who sets it:** FDA regulations, including **21 CFR Parts 210 and 211** for drugs, additional parts for biologics, and quality system rules for medical devices. International guidance comes from ICH and other bodies.

**What it requires, in outline:**

- A **quality unit** with authority to approve or reject materials and batches.
- **Batch records** documenting every step of making each batch.
- **Validated** processes, equipment, methods and cleaning.
- **Change control**: no change to a process without review and approval.
- **Deviations** investigated and **CAPA** (corrective and preventive action) to stop them recurring.
- Training records for every employee.

A biological technician at a vaccine plant or a QC analyst at a pharma company lives in GMP every day.

## CLIA: Clinical Laboratory Improvement Amendments

**What it covers:** US laboratories that test **human specimens** to provide information for the diagnosis, prevention or treatment of disease, or the assessment of health. Hospital labs, reference labs and doctor's office labs fall under CLIA. Research labs that do not report patient-specific results for patient care are generally not covered.

**Who runs it:** **CMS** (Centers for Medicare & Medicaid Services) administers CLIA; **FDA** categorises tests by complexity; **CDC** provides scientific support. Labs hold a CLIA certificate, and many are accredited by an approved organisation such as **CAP** (College of American Pathologists).

**What it requires, in outline:**

- Tests classified as **waived**, **moderate complexity** or **high complexity**, with stricter rules as complexity rises.
- **Quality control** for every test run and **proficiency testing** (blind samples from an outside program).
- **Personnel qualifications** for directors, supervisors and testing staff, which is one reason clinical employers require MLS(ASCP)-level training.
- Inspections, and corrective action when problems are found.

## Which rules apply? A worked example

| Situation | Main rule set |
|---|---|
| A toxicology study in rats to support a new drug application | GLP |
| Filling vials of a vaccine and testing each batch before release | GMP |
| A hospital measuring a patient's potassium level | CLIA |
| A university lab studying how yeast cells divide | None of the three (institutional safety and ethics rules still apply) |
| A clinical trial giving a new drug to patients | GCP (Good Clinical Practice), plus CLIA for any lab testing that supports patient care |

Official sources for detail: FDA ([fda.gov](https://www.fda.gov)), CMS CLIA pages ([cms.gov](https://www.cms.gov)) and the CDC's CLIA resources ([cdc.gov](https://www.cdc.gov)). Regulations and guidance change; read the current versions there. Checked on 2026-10-23.

## Common mistakes

- **Using "GLP" to mean "good lab habits".** In regulated work, GLP is a specific legal framework for safety studies.
- **Assuming research labs follow GMP.** Most academic research does not, which is why industry trains new hires in GMP.
- **Thinking CLIA is only about equipment.** It also covers staff qualifications, QC and proficiency testing.

## Quick recap

- GLP: integrity of non-clinical safety studies submitted to regulators (21 CFR 58).
- GMP: consistent, safe manufacturing and batch release (21 CFR 210/211 for drugs and more).
- CLIA: quality of patient testing on human specimens, run by CMS with FDA and CDC.
- Each requires written procedures, trained staff and records an inspector can check.


Explain which US rules govern which labs: GLP for safety studies, GMP for manufacturing, CLIA for clinical testing.

GLP, GMP, and CLIA

GMP vs GLP vs GCP Difference in 10 Minutes | FDA Compliance Guide

Understanding CLIA and CAP Regulations to Advance Your Laboratory Career

## Why it matters

Life science research often involves people (blood samples, medical records, clinical trials) or animals. History shows what happens when researchers decide for themselves what is acceptable. Today, research involving humans or animals must be reviewed and approved before it starts, and everyone on the team, including technicians and students, must follow the approved protocol. You will complete ethics training before working on such projects; this lesson gives you the background.

## Lessons from history

- **The Nuremberg Code (1947)** came out of the trials of Nazi doctors who experimented on prisoners. Its first principle: the voluntary consent of the human subject is absolutely essential.
- **The US Public Health Service syphilis study at Tuskegee (1932-1972)** followed hundreds of Black men with syphilis without telling them their diagnosis, and withheld effective treatment (penicillin) even after it became standard. Public exposure in 1972 ended the study and led directly to US research protections.
- **Henrietta Lacks (1951).** Cells taken from her tumour at Johns Hopkins without her knowledge became the HeLa cell line, used in countless discoveries. Her family was not told for decades. The case still shapes discussion of consent, privacy and the use of biological samples.

## The Belmont Report's three principles

In 1979 a US national commission published the **Belmont Report**, which still underpins research ethics:

1. **Respect for persons:** treat people as autonomous; get informed consent; protect those with reduced autonomy (children, people with cognitive impairments, prisoners).
2. **Beneficence:** do no harm, and maximise benefits while minimising risks.
3. **Justice:** distribute the burdens and benefits of research fairly; do not target vulnerable groups because they are convenient.

## IRBs and the Common Rule

In the US, most federally funded research with human subjects follows the **Common Rule** (45 CFR 46), and FDA-regulated studies follow FDA's human subject rules. Both require review by an **Institutional Review Board (IRB)**: a committee including scientists, non-scientists and at least one member not affiliated with the institution. The IRB:

- Reviews the protocol before research begins.
- Weighs risks against benefits.
- Checks the consent process and documents.
- Can require changes, approve, or disapprove, and continues oversight while the study runs.

Some low-risk studies qualify for **exempt** or **expedited** review, but the decision belongs to the IRB, not the researcher.

## Informed consent

Consent is a **process**, not just a signature. Participants must be told, in language they understand:

- The purpose, procedures and how long participation lasts.
- Foreseeable risks and discomforts, and possible benefits.
- Alternatives, and how their data and samples will be kept confidential and used.
- That participation is **voluntary** and they can **withdraw at any time** without penalty.
- Whom to contact with questions.

## Animal research: IACUCs and the 3Rs

Research with vertebrate animals at US institutions is reviewed by an **Institutional Animal Care and Use Committee (IACUC)** under the Animal Welfare Act and Public Health Service policy. Researchers must justify the use of animals and follow the **3Rs**:

- **Replace** animals with alternatives (cell culture, computer models) where possible.
- **Reduce** the number of animals to the minimum needed for a valid result.
- **Refine** procedures to minimise pain and distress.

## Research misconduct

Separate from consent, US agencies define research misconduct as **fabrication** (making up data), **falsification** (manipulating data or images) and **plagiarism**. These end careers. Honest error and differences of opinion are not misconduct; hiding errors can become it.

## A worked example

A technician in a lab studying a blood biomarker is asked to "just run a few extra samples" from the freezer for a new side project. The samples were collected under a consent form that covered only the original study. What should happen?

The technician should pause and ask whether the IRB-approved protocol and the consent cover the new use. If not, the PI needs to seek IRB approval (and possibly new consent) first. Doing the work without it would breach respect for persons, regardless of how harmless it seems.

## Quick recap

- History (Nuremberg, Tuskegee, HeLa) explains why research protections exist.
- Belmont principles: respect for persons, beneficence, justice.
- IRBs review human research before it starts; consent is an informed, voluntary, ongoing process.
- IACUCs review animal research using the 3Rs; fabrication, falsification and plagiarism are misconduct.
- Official sources: HHS Office for Human Research Protections ([hhs.gov/ohrp](https://www.hhs.gov/ohrp)) and NIH Office of Laboratory Animal Welfare ([olaw.nih.gov](https://olaw.nih.gov)). Checked on 2026-10-23.


Explain how IRBs, informed consent and animal welfare committees protect people and animals in research.

Research ethics and consent

How IRBs Protect Human Research Participants

What is informed consent?

## Why it matters

Every decision a lab supports (a patient's diagnosis, the release of a vaccine batch, a published discovery) is only as good as the records behind it. Regulators found that many serious problems were not bad science but bad records: missing data, backdated entries, results re-run until they passed. **Data integrity** means data are complete, consistent and accurate throughout their life. The **ALCOA+** principles are the shared checklist regulated labs use, and they are good practice in any lab.

## ALCOA

The original five, popularised at the US FDA:

- **Attributable:** you can tell who recorded the data and who did the work, and when. Initials or signatures on paper; unique user logins (never shared passwords) on instruments and software.
- **Legible:** readable and permanent for as long as the record must be kept. Ink, not pencil; no erasing; electronic files in formats that can still be opened.
- **Contemporaneous:** recorded at the time the work was done, not later from memory or scrap paper.
- **Original:** the first record (or a verified "true copy"). The printout from the balance, the raw instrument file, the notebook entry. Transcribed numbers are not the original.
- **Accurate:** correct, with no errors or editing that is not documented. Changes are recorded with who, when and why, leaving the original visible.

## The "+"

Later guidance added four more:

- **Complete:** all data, including repeats, failed runs and re-tests, are kept. Nothing is deleted or left out.
- **Consistent:** entries follow the expected sequence, with dates and times in order and the same units and formats.
- **Enduring:** records last for the required retention period (bound notebooks, backed-up systems).
- **Available:** records can be found and read when needed, for review, audit or inspection.

Some organisations use **ALCOA++**, adding traceability, but the core idea is the same.

## Electronic records

Most lab data now live in instruments, LIMS and electronic notebooks. In the US, FDA rules on electronic records and signatures (**21 CFR Part 11**) apply to many regulated systems. Integrity features include unique logins, **audit trails** that log every create, change and delete with who, when and why, and controlled backups. A common inspection finding is a shared "lab" login, which makes every result non-attributable.

## A worked example: audit a record

Here is a QC record for a pH check of a buffer:

```text
Buffer lot B-0912   pH spec 7.30-7.50
Result: 7.41        Checked by: Lab        Date: (blank)
Note on sticky note attached: "1st reading 7.62 - recalibrated, re-ran"
```

Check it against ALCOA+:

| Principle | Problem |
|---|---|
| Attributable | "Lab" is not a person. Who measured it? |
| Contemporaneous / Consistent | No date or time, so you cannot place it in sequence. |
| Original | The first reading lives on a loose sticky note, not in the record. |
| Complete | The out-of-spec reading (7.62) is not part of the official record. |
| Accurate | Was the meter calibrated before the first reading? No record. |

A compliant version:

```text
2026-10-23 09:12  Buffer lot B-0912, spec pH 7.30-7.50.
pH meter #2, calibration check 09:05 with pH 7.00/10.00 buffers: FAILED slope.
Recalibrated 09:08 (pass, slope 98.7%).  Reading 1 (before recal) 7.62 - invalid,
  meter failed calibration check; deviation DEV-0331 opened.
Reading 2 09:12: 7.41  (PASS)                         -- A. Patel (AP)
Reviewed 2026-10-23 11:30 -- R. Gomez (RG)
```

Every value is kept, every action is attributable and dated, and the reason the first reading was invalid is documented and linked to a deviation. In a GMP lab, simply re-testing until a result passes ("testing into compliance") without an investigation is a serious violation.

## Common mistakes

- **Shared logins** or signing for someone else.
- **Writing on scrap paper** and copying later.
- **Deleting or not recording failed runs.**
- **Backdating** an entry to "fill a gap". Late entries must be marked as late, with the actual date written.
- **Unreviewed changes** to electronic data with the audit trail turned off.

## Quick recap

- ALCOA: Attributable, Legible, Contemporaneous, Original, Accurate.
- Plus: Complete, Consistent, Enduring, Available.
- Keep every result, including failures; document every change with who, when and why.
- Electronic systems need unique logins and audit trails (21 CFR Part 11 in many US regulated labs).
- FDA guidance on data integrity is at [fda.gov](https://www.fda.gov); read the current version there. Checked on 2026-10-23.


Apply the ALCOA+ principles to judge whether a lab record is trustworthy, and fix one that is not.

Data integrity (ALCOA+)

ALCOA+ - An easy abbreviation for data integrity

ALCOA and ALCOA+ in Pharmaceuticals | Principles of ALCOA | Data Integrity Principles

GLP, GMP and CLIA, research ethics and consent, and ALCOA+ data integrity.