“Questions of Quality” is 30 years old! What, if anything, has changed in chromatography laboratories over that time? What, if anything, remains the same?
Back to the future! The origins of this column lie at the 1993 Pittsburgh Conference, when I was asked by Paul Moss, then editor of LCGC International, to write a six-part series on—wait for it—data integrity. The first column was published in September (1), and the original theme of data integrity has been the subject of many columns since then. Let’s start with a summary of what I wrote 30 years ago and see what, if anything, has changed.
The aim of the first column was to consider data integrity within an automated chromatography laboratory (1). My key points were:
Readers of this column will see that these are recurrent themes that I have written about in later instalments. Recently, I discussed how laboratories are still printing paper instead of working electronically (2)—some chromatographers never learn.
References to outdated technology such as optical disks and binary-coded decimal (BCD) communication of vial numbers to the chromatography data system (CDS) have not stood the test of time. This only goes to show that you can’t win them all.
Two of my past columns have discussed sample labelling using barcodes, with the suggestion of looking at your local supermarket to see how they use barcodes to manage stock and then taking these principles and applying them in your own laboratory (3,4). There was practical advice to ensure that you had the right tack label adhesive for the type of container and storage conditions. This was an example of Horner’s Five Thumbed Postulate in action (the amount of experience is directly proportional to the number of systems ruined), as I had bought a quarter of a million sample labels with the wrong tack adhesive.
There were columns on how to evaluate and purchase a chromatograph and a CDS. In the former there was mention of a cupboard that all laboratories have, typically furthest from the head of the lab’s office, where they stash all instrument disasters (5). If your laboratory suffers from an end of fiscal year slush fund spend where you have a month to use it or lose it, you’ll find plenty of candidates for the cupboard. At least you have a fighting chance of hiding your mistakes, unlike a CDS where the wrong system is visible with no hiding place (6). Writing a specification is critical and varies from minimal for a chromatograph to more extensive for a CDS (7).
In an early column I discussed why kitchen designers have a better understanding of layout and the need to integrate the sink, cooker, and food preparation areas than the average chromatographer has in laboratory design (8). My evidence was based on the fact that laboratory benches have been unchanged since 1609 (the year not the time) as shown in an antique print. Alternative ways of designing laboratories were presented, including putting chromatographs on moveable trolleys with overhead services or having a space behind a single bench for easy access to all parts of an instrument.
CDS applications are much improved since 1993. An early column for evaluation of a CDS mentioned the lack of calibration models and the inability to perform all post-run calculations (6), which are now much improved. CDS architecture has changed from files stored in operating system directories or folders to databases, which has resulted in much improved data integrity functions and access to data across all chromatographs. There are functions to control different supplier’s chromatographs, electronic signature capability, and interfacing with LIMS to transfer sample identities, weights, and individual and reportable results between the two applications.
However, chromatographers are still buying standalone rather than networked systems because the former suffer from a number of issues such as keyboard contention, inability to share data easily— leaving chromatographers responsible for backup and recovery—and suffer a lack of resilience with single points of failure. Networked systems are much better. Standalone systems tend to be used in hybrid mode, with lax technical controls and potential conflicts of interest. In many cases, raw data or complete data (9,10) are defined as paper printouts, which is most certainly not the case, as it is the electronic record that is the foundation of raw data or complete data (11,12). Networked systems that are designed to work electronically provide both regulatory compliance and business efficiencies.
The CDS is only half the story. As stated in the very first column, the process must be considered when automating. In 2005, Jens Donath and I published a case study about the validation of a site-wide CDS (13). The key to success was the mapping of the current (As-Is) process and the identification of bottlenecks and improvement ideas, which were used to develop the future (To-Be) process. This was achieved by two two-day workshops, with a report describing each process and highlighting the process improvements. The aim of the redesign was to have an electronic process with signatures that met the requirements of both good manufacturing practice (GMP) (14,15) and 21 CFR 11 (13). The CDS was then implemented and validated based on the To-Be process. Management is critical to supporting such an initiative, but the payback on the investment is considerable. Twenty years on, the same unchanged process is implemented across all six business units and all sites within the organization (16), which shows the power of process design and getting it right first time. Incidentally, screenshots to support the validation were kept to a minimum.
Since the Able Laboratories fraud case in 2005 involving a CDS (17), regulators have found more CDS data integrity problems, such as test injections, deletion of data, and integration into compliance. This culminated in a column instalment on the role of a CDS in falsification and fraud, with 10 suggestions made to avoid data integrity problems (18). Peak integration is a major way of ensuring compliance with specifications, but is also a subtle way to falsify results; tips to manage the process featured in two columns in 2015 and 2019 (19,20). Historically, 25% of data integrity warning letters were integration related (21)—so control peak integration.
Always remember that the role of a CDS is not to correct your crummy chromatography. Good peak shapes result in good integration.
Allied to process redesign is my war on spreadsheets. As a chromatographer and a consultant, get rid of them! As an auditor, this is a gift that keeps on giving, so please keep using them.
Why use a spreadsheet? A CDS can perform calculations without the need to print the peak areas and manually enter the values into a spreadsheet. The number of instances where a CDS is only used as an electronic ruler is amazing, even today. It appears that the lure of the spreadsheet is greater than the common-sense approach of using a CDS to perform calculations, such as system suitability test (SST) parameters and sample results.
To illustrate the silliness of this approach, a 2020 “Questions of Quality” column (22) grew from an FDA warning letter (23) that cited an unvalidated spreadsheet where the inspector had found a calculation error. This was compounded during remediation by the company finding another error!
by the company finding another error! I presented a current process where the CDS was an electronic ruler and two spreadsheets, one for SST calculations and one for the results calculations, were used. The process featured a stellar cast of characters, including Dopey, the head of quality control (QC), and Stupid from quality assurance (QA) (22), who between them allowed such an abysmal process to be established. This is not a fairy tale and names were changed to protect the guilty.
Having the right instrument for the job is a key requirement for data integrity, and requires qualification of each chromatograph. Since the publication of United States Pharmacopeia (USP) <1058> on Analytical Instrument Qualification (AIQ) in 2008 (24), discussion and interpretation of the approach has featured in this column with co-authors Chris Burgess or Paul Smith (25–28). The current version of the general chapter integrates instrument qualification with computerized system validation (CSV) (29). An update to USP <1058> is scheduled soon, with a name change to Analytical Instrument and System Qualification (AISQ). Articles outlining some possible changes to the general chapter have been published (30–33).
Analytical balances are at the heart of any chromatographic analysis, and the principles of operation must be understood and controlled to ensure the foundation of reliable and trustworthy results. Chris Burgess and I wrote two columns looking at the principles of an analytical balance and how to apply those principles in practice (34,35).
For 30 years regulators have tried to persuade or cajole laboratories to control blank forms for recording analytical work (36–39). If results don’t meet expectations, the original record can be replaced by a new form straight from the photocopier and the analysis repeated. This is not compliant and is actually falsification. Chris Burgess and I wrote an article on control of the master templates and blank forms (40). There is a high administrative overhead to manage this; it is much better to work electronically and eliminate paper.
LIMS have featured in “Questions of Quality” over the years. The first two articles explained what a LIMS is and the 10 steps to heaven, sorry, to effective LIMS implementation (41,42). As a result of data integrity problems, I also wrote a column on how a LIMS can help ensure data integrity (43).
Regulatory documents from the World Health Organization (WHO), Indian Pharmacopoeia, and the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) have been on the receiving end of extremely critical comments by both myself and in collaboration with Chris Burgess. I described the WHO guidance for good chromatography practices (44) as having sections organized at random with no logical order, and document numbering not following a numerical sequence (45,46). It should be fast-tracked to the round filing cabinet under your desk or the recycle bin.
Chris Burgess and I reviewed the Indian Pharmacopoeia’s general chapter for qualification of high performance liquid chromatography (HPLC) instruments as unscientific dross— enough said (47). We also reviewed the draft ICH Q2(R2) and ICH Q14 for validation of analytical procedures (48,49). The original aim of the working party was to write a single document but they failed; it is not an integrated approach to development, validation, and operation of analytical procedures (50). Use USP <1220> instead for a better integrated approach (51).
My philosophy when writing “Questions of Quality” has been to present a topic in simple terms that is intended to stimulate you to think. I’m not interested in whether you agree or disagree with me, I want you to think and come to your own conclusions. Quality, integrity, and compliance may be the world’s second most boring subject, so I try to inject a degree of humour into my writing, although scepticism and sarcasm also appear.
The trigger for starting this column was data integrity, and throughout the 30 years of this column the quality and integrity of data have been my focus. With the major problems of data integrity in the pharmaceutical industry since 2005, what goes around comes around.
Let me answer the questions posed at the start of this column:
What has changed?
What remains the same?
I would like to thank all my co-authors and reviewers of all “Questions of Quality” columns over the past 30 years, it has been very satisfying and enjoyable collaborating with you all. Thanks to Chris Burgess, Paul Smith, and Mahboubeh Lotfinia for constructive review of this column.
(1) McDowall, R. D. Questions of Quality: Quality or be Damned! LCGC International 1993, 6 (9), 550–552.
(2) McDowall, R. D. The Hidden Factory in Your Laboratory? LCGC Eur. 2021, 34 (8), 326–330.
(3) McDowall, R. D. Questions of Quality: I Can’t Read the Label! LCGC International 1993, 6 (10), 612–615.
(4) McDowall, R. D. Questions of Quality: Bar Codes: Sample Labels and Beyond. LCGC International 1993, 6 (11), 690–691.
(5) McDowall, R. D. Questions of Quality: The Right Tool for the Right Job? LCGC International 1994, 7 (3), 146–148.
(6) McDowall, R. D. Questions of Quality: Evaluating a Chromatography Data System. LCGC International 1994, 7 (5), 258–265.
(7) McDowall, R. D. How Do You Write User Requirements for Chromatographs and Chromatography Data Systems? LCGC Eur. 2020, 33 (5), 257–263.
(8) McDowall, R. D. The Alchemist’s Laboratory. LCGC International 1994, 7 (1), 16–19.
(9) McDowall, R. D. How Complete Are Your Chromatographic Data? Part 1. LCGC Eur. 2013, 26 (6), 338–343.
(10) McDowall, R. D. How Complete Are Your Chromatographic Data? Part 2. LCGC Eur. 2013, 26 (7), 389–392.
(11) FDA, Questions and Answers on Current Good Manufacturing Practices, Good Guidance Practices, Level 2 Guidance - Records and Reports (December 2019).
(12) Burgess, C.; McDowall, R. D. What’s in a Name? LCGC Eur. 2015, 28 (11), 621–626.
(13) Donath, J.; McDowall, R. D. Cost Beneficial Validation of a Site-Wide Chromatography Data System. LCGC Eur. 2005, 19 (9), 453–464.
(14) FDA, 21 CFR 211 Current Good Manufacturing Practice for Finished Pharmaceutical Products (2008).
(15) EudraLex, Volume 4 Good Manufacturing Practice (GMP) Guidelines, Annex 11 Computerised Systems (European Commission, 2011).
(16) Author’s personal communication
(17) Able Laboratories Form 483 Observations, 2005. https://www.fda.gov/media/70711/download (accessed 2023-08-10)
(18) McDowall, R. D. The Role of Chromatography Data Systems in Fraud and Falsification. LCGC Eur. 2014, 27 (9), 486–492.
(19) McDowall, R. D. Where Can I Draw The Line? LCGC Eur. 2015, 28 (6), 336–342.
(20) Longden, H.; McDowall, R. D. Can We Continue to Draw the Line? LCGC Eur. 2019, 21 (12), 641–651.
(21) Author’s personal communication
(22) McDowall, R. D. Are Spreadsheets a Fast Track to Regulatory Non-Compliance? LCGC Eur. 2020, 33 (9), 468–476.
(23) FDA Warning Letter Tismore Health and Wellness Pty Limited, 2019. https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/tismor-health-and-wellness-pty-limited-588104-12052019 (accessed 2023-08-10).
(24) USP, USP 31 General Chapter <1058>, “Analytical Instrument Qualification,” (Rockville, Maryland, USA, 2008).
(25) Smith, P.; McDowall, R. D. Life Cycle Risk Assessment of HPLC Instruments. LCGC Eur. 2015, 28 (2), 110–117.
(26) Smith, P.; McDowall, R. D. Data Integrity and USP <1058>: Part 1 Specifications and Suppliers. LCGC Eur. 2018, 31 (7), 385–389.
(27) Smith, P.; McDowall, R. D. Data Integrity and USP <1058>, Part 2: OQ Supervision and Execution. LCGC Eur. 2018, 31 (9), 504–511.
(28) Smith, P.; McDowall, R. D. Data Integrity and USP <1058>: Part 3: Monitoring and Requalification. LCGC Eur. 2019, 32 (1), 28–32.
(29) USP, USP General Chapter <1058>, “Analytical Instrument Qualification,” (Rockville, Maryland, USA, 2008).
(30) Weitzel, M. L. J. et al., “Measurement Uncertainty Evaluation Relevant to Analytical Instrument and System (AIS) Qualification - The Role of Measurement Uncertainty Concepts within the AIS,” Pharmacopoeial Forum 48 (2) (2022).
(31) Burgess, C.; Weitzel, M. L. J.; Roussel, J.-M.; et al., “Analytical Instrument and System (AIS) Qualifcation to Support Analytical Procedure Validation over the Life Cycle,” Pharmacopoeial Forum 48 (1) (2022).
(32) Burgess, C.; et al., “Analytical Instrument and System (AIS) Qualification: The Qualification Life Cycle Process,” Pharmacopoeial Forum 48 (2022).
(33) Burgess, C.; McDowall, R. D. Quo Vadis Analytical Procedure Development and Validation? Spectrosc. 2022, 37 (9), 8–14.
(34) Burgess, C.; McDowall, R. D. A Question of Balance? Part 1: Principles. LCGC Eur. 2004, 17 (7), 390–395.
(35) Burgess, C.; McDowall, R. D. A Question of Balance? Part 2: Putting Principles into Practice. LCGC Eur. 2006, 19, 152–159.
(36) FDA, Inspection of Pharmaceutical Quality Control Laboratories (1993).
(37) WHO, WHO Technical Report Series, “No.996 Annex 5 Guidance on Good Data and Records Management Practices,” (Geneva, Switzerland, 2016).
(38) FDA, Guidance for Industry Data Integrity and Compliance With Drug CGMP Questions and Answers (2018).
(39) PIC/S, PI-041 Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments (2021).
(40) Burgess, C.; McDowall, R. D. Paper, Paper Everywhere but None of it Controlled. LCGC Eur. 2016, 29 (9), 498–504.
(41) McDowall, R. D. Questions of Quality: Not Another LIMS Project? LCGC International 1996, 9 (1), 21–26.
(42) McDowall, R. D. Questions of Quality: Ten Steps to Heaven? LCGC International 1996, 9 (4), 202–213.
(43) McDowall, R. D. How Can LIMS Help Ensure Data Integrity? LCGC Eur. 2016, 29 (6), 310–316.
(44) WHO, WHO Technical Report Series, “No.1025, Annex 4 Good Chromatography Practices,” (Geneva, Switzerland, 2020).
(45) McDowall, R. D. What’s Good About the WHO Good Chromatography Practices Guidance? Part 1. LCGC Eur. 2020, 33 (11), 579–584.
(46) McDowall, R. D. What’s Good About the WHO Good Chromatography Practices Guidance? Part 2. LCGC Eur. 2021, 34 (2), 66–71.
(47) Burgess, C.; McDowall, R. D. It’s Qualification, But Not As We Know It? LCGC Eur. 2021, 34 (11), 476–481.
(48) ICH, Q2(R2) Validation of Analytical Procedures, Step 2 draft (2022).
(49) ICH, Q14 Analytical Procedure Development. Step 2 draft (2022).
(50) Burgess, C.; McDowall, R. D. Quo Vadis Analytical Procedure Development and Validation? LCGC Eur. 2022, 35 (7), 279–284.
(51) USP, USP General Chapter <1220>, “Analytical Procedure Lifecycle,” (Rockville, Maryland, USA, 2022).
(52) Burgess, C.; McDowall, R. D. The Ideal Chromatography Data System for a Regulated Laboratory, Part 1: The Compliant Analytical Process. LCGC N. Am. 2015, 33 (8), 554–557.
(53) Burgess, C.; McDowall, R. D. The Ideal Chromatography Data System for a Regulated Laboratory, Part 2: System Architecture Requirements. LCGC N. Am. 2015, 33 (10), 782–785.
(54) Burgess, C.; McDowall, R. D. The Ideal Chromatography Data System for a Regulated Laboratory, Part 3: Essential Chromatographic Functions for Electronic Ways of Working. LCGC N. Am. 2015, 33 (12), 914–917.
(55) Burgess, C.; McDowall, R. D. The Ideal Chromatography Data System for a Regulated Laboratory, Part 4: Assuring Regulatory Compliance. LCGC N. Am. 2016, 34 (2), 144–149.
Bob McDowall is Director of R.D. McDowall Ltd., Bromley, UK. He is also a member of LCGC Europe’s editorial advisory board. Direct correspondence to: amatheson@mjhlifesciences.com