Thanks for coming back for part II! In part I, I gave a bird’s-eye overview of our phage therapy service at Phage Australia.
In this post I’ll get into some of the ways we’ve improved our efficiency as we select and make phages for patients, such as dividing up work while ensuring nothing depends on just one person, managing expectations among clinical and informatics teams, and how we’re setting up lab documentation processes to support it all.
How we balance division of labour with redundancy
Our main goal is being able to respond to patient cases that come up, without overwhelming our small team or compromising patient safety. But this is hard, and requires a system.
Luckily, for the work that happens in the phage lab, there are two of us phage biologists: myself and Stephanie. From an efficiency perspective, it’s become important for us to have sections we each ‘own’ (much faster for decision making, and better for focus). So we’ve gradually divided up our work. Currently, Steph does diagnostics, I do production and purification, and she does quality control and monitoring.
However, the more we divide up work, the more at-risk our process becomes to delays/failure if one of us is away, sick or busy.
To manage this, we spend a lot of time making sure we can both dip in to cover either person’s section on a moment’s notice. To do this, we rely on establishing clear protocols and making sure we are each able to make them work, even if they’re ‘in the other person’s section’.
We’re very lucky there are two of us and we’re both phage biologists, and thus we do have some redundancy.
In other categories, we find other ways to establish redundancy. For example, if our bioinformatician, Nouri (who verifies safety of our phages and strains as we test and make them), is away, we do have a hospital sequencing and bioinformatics service we can turn to in a pinch.
My overarching point is we’re working on a balance between division of labour (which makes sense with a team of specialists and many categories of work that have to happen on time) with having redundancy and contingency plans wherever possible. It’s a work in progress, but it’s probably the most important thing that makes it possible for us to keep responding to patient cases as they come in.
How we hand work from one person to the next
For redundancy purposes, and for complex choreography of multiple domains (lab, genomics, clinic), we’ve found it crucial to think intentionally about how we hand work over from one person to the next.
Our handoffs originally involved a lot of verbal back and forth. We’re now getting to a point where our handoffs have more defined thresholds to meet, and more of a formalized paper trail. For example Stephanie may identify three phages for patient X. She’ll now give me vials of each of those phages and a phage susceptibility report, and that’s how I’ll know which phages to purify and what their production strains are. Or for QC handoff, once I finish producing a phage to an appropriate titre and endotoxin level, I hand Stephanie a vial of concentrated, purified phage, plus a draft batch report to signify the name of the phage, the route of administration (which impacts our QC thresholds), and give her a place to fill in the QC values she collects.
This is another area to balance, because too much paperwork feels like it slows us down, but too much verbal conversation leaves details lost to the ether, and relies too much on our memory. But we’ve at least identified handoffs as a crucial area that makes the difference between a smooth, functioning process and one that falls apart.
How we manage expectations without compromising on safety
Managing expectations has been another tricky area. Doing all the steps we do to select and produce phages isn’t too hard if you only have one patient, and as much time as it takes. But once you’ve got many patients, and people (mainly on the clinical team, but also couriers and bioinformaticians) who need to know exactly which day the phage will be ready, complexity increases and efficiency can screech to a halt.
We’re still struggling with this, but we’ve started establishing realistic timeframes and boundaries for ourselves, and communicating them with all the different groups we work with.
For example, we currently accept new patient bacterial strains on a rolling basis for diagnostics (since this part of our process is fairly quick, batchable, and gives us data about our phages even if a patient case doesn’t go ahead). However, production is another story; more work, less batchable. So we now only choose one patient at a time to enter into the production phase, AND we ensure that we only move a patient case forward to this phase if their hospital/clinical team has all their trial paperwork filled out.
In terms of setting timelines, we now tell clinical teams we need 4-8 weeks for production, QC and formulation, even though it’s technically only about 2 weeks of work if all goes well. 4-8 weeks tends to cover unexpected happenings (since of course nothing ever all goes to plan…), and we’ve found that having more predictability ends up leading to an overall more efficient process, even if it’s not as fast.
Ultimately it’s always tempting to give short timelines, be flexible, and make exceptions for patients in need, but overall we’re much more efficient when we have established timeframes and rules for ourselves, hold firm on what we’ve established, and overcommunicate it all.
The power of recurring, function-specific meetings
It helped us a lot once we started a weekly clinical triage meeting; this happens every Thursday and is the only time new phages can be assigned for production, and new cases can be added for diagnostics. Having a specific time when we get new work has helped with overwhelm.
We also have recurring meetings like diagnostics/production and genomics/informatics, where we discuss issues with protocols or required improvements.
For sharing results, we have larger, separate, more open meetings, like our ‘G-CPR: International Clinical Grand Rounds’, where patient outcomes and experiences are discussed by clinicians and scientists who work with them. We also have Phage Australia lab meetings, which is where we showcase progress and get help from our fellow Phage Australia consortium members.
Our version-controlled SOP + log system
Creating a set of standard operating procedures (SOPs) has been a game changer. To accommodate the redundancy I described above, and to fulfill drug committee reporting requirements, our phage therapy service-related protocols need to be agreed upon and well defined. And since our protocols are shifting a lot as we improve them, we needed to create a versioning system to accommodate new versions without overwriting the old ones.
I’ll write more about this in a future post, but briefly, we organize our SOPs into different folders (each dedicated to a category of work, like diagnostics, production, purification, QC, etc). We keep all these in a shared drive, and have a single current version of each in the main folder for that category. When we create a new version, we make a copy of the current one, change the version number, make changes, and then we drag the one we copied into the ‘Archived’ folder for that category.
We’ve also found that we need ‘logs’ for each SOP we have: each time we run a protocol, we need to record certain things each time. This is especially helpful so we can look back if something looks weird to see if anything might explain it. Especially since we have multiple people sometimes doing the same piece of work, it helps to be able to see any steps that were done differently, or to flag anything weird that happened that time compared to previous times.
In summary, I hope I’ve illustrated through this (so far 2-part) series, a bird’s eye view of how our ‘personalized phage therapy as a service’ system currently works at Phage Australia. With our ‘minimum viable pipeline’ for diagnostics, production, QC and monitoring (check out part I for a refresher), we are now successfully treating one patient per month with phages we produce in-house.
Our latest efficiency unlocks have come from deciding on thresholds to hold ourselves to (minimum safety standards, minimum doses to achieve), dividing up work and establishing redundancy and clear handoff points, coming up with boundaries and timelines for our work, and continually communicating these to the teams across the various disciplines we rely on (e.g. clinical, informatics).
Our next frontier is establishing our lab documentation system, starting with nailing down standard operating procedures and corresponding logs for each step of our process. We’re actively working on this, and I hope to write more about this soon!
- A talk I gave at Targeting Phage Therapy 2023 (Paris, France): Helping scale phage therapy, from Phage Directory to Phage Australia
- Labspace.ai: GPT-based tools for automating lab work, made by Jan and me; includes our Phage Australia vial labelmaker
- My blog posts on our Phage Australia processes & journey (2022-2023):
- Slides from a talk I gave on our Phage Australia process at the 2022 Bioprocessing Network Conference (Sydney, Australia)
- A poster I presented on our phage therapy pipeline at the 2022 Australasian Virology Society conference (Gold Coast, Australia)
- STAMP Study: Khatami, A. et al. (2022). Standardised treatment and monitoring protocol to assess safety and tolerability of bacteriophage therapy for adult and paediatric patients (STAMP study): protocol for an open-label, single-arm trial. BMJ open, 12(12), e065401.
- Phage Australia website: https://phageaustralia.org
- Paper by Khatami et al 2021: Bacterial lysis, autophagy and innate immune responses during adjunctive phage therapy in a child
- Paper by Petrovic Fabijan et al 2020: Safety of bacteriophage therapy in severe Staphylococcus aureus infection
- We didn’t come up with this system from scratch! It was inspired by:
- Phage physician desk reference (TAILOR Labs at Baylor College of Medicine)
- Queen Astrid Military Hospital (clinical triage flow)
- Sciensano (phage batch QC method; Magistral Phage)