Start here

Normalisation and the wheel and uniformity

With any recommended introduction or change of a practice, one needs to consider how it is to be implemented, they’ve even dedicated a science to it, a journal the whole lot! One group I have enjoyed reading is Carl Mays group and their normlisation process theory For somebody new to this area, the NPT seems to be quite practical, and importantly relevant and applicable to my work. More of this to come in the future no doubt.

Have been reading some interviews conducted with some infection prevention practitioners (for a colleague) and have come across some interesting comments. Possibly not original, but worth a thought, and actually, I think they are really relevant to my work (which I guess is why I remember them)…

“Everybody says dont re-invent the wheel. Well I think we should. If we didnt re-invent the wheel, we’d still have wooden wheels”

Well yeah, I can see their point. Sometimes the answer is there right in front of you, its just needs a tweek. In fact, you are not really re-inventing the wheel, you’re improving the current design…but I’m probably being a bit technical. It’s akin to the notion of “leveraging off existing systems”. Sometimes you do have to look at what is there, and and ask is it fit for purpose, or does it need a complete redesign? On this point, Australia already has a national surveillance program, it just doesn’t include HAIs. The Communicable Diseases Network of Australia (CDNA) has been in existence now for a few decades and collates data from the States and Territories on notifiable infectious diseases. It has direct links to the Chief Medical Officer, has widespread jurisdictional buy in, and expertise in surveillance. Although not all States and Territories have identical lists of notifiable diseases, where feasible, CDNA collates data and produces national figures. This gives us an understanding of the epidemiology of these diseases in Australia. So why don’t we just tell them to start collating data on HAIs? Well I guess they could, though with the variation of practices and poor agreement already identified, the data collated would have so many footnotes and  conditions attached to it, we know it wouldn’t be worth doing. But lets just say, one day in the future, we are all doing HAI surveillance the same way, could they collate and publish it then? Most likely, with a few resources dedicated to HAI (is this starting to sound like an Australian version of CDC? Imagine!) we could have embedded a national HAI program into am already exiting structure (the wooden wheel?). But what then? With the existing CDNA, if a notifiable disease is identified it is up to the relevant State or Territory department to manage the situation, launch an outbreak investigation, inspect kitchens, notify schools, whatever. But what happens if a concern (an increase) with a HAI is identified? Who is going to look into it then? Who is the “action” arm of this process? And that’s where it all falls down. Not all of our State or Territories have an action arm. And of course, whats the point of surveillance if there is no action…

So, simply put, with no action arm, why bother collecting the data. If we are ever going to have a national surveillance program, we must not only consider the process of collecting, analysing and feeding back data, we need to also consider the who is going to govern the interventions. If its not data for action, then don’t bother.

Melbourne winters, a paper and a DCE

Such is the size of Australia that if you don’t like the weather you just travel to a better climate without a passport. In Melbourne we are approaching the end of the dark depths of winter, with recent temperatures maxing at 10C, with a “feels like” factor of 5C. I don’t quite get the “feels like” measure, but I understand it takes in a few criteria, including heat loss from the body. Nevertheless, its been cold, and several lucky colleagues have escaped to warmer climes up north. Spring is only days away!

Happy to provide the link of my second paper from my first study which describes agreement levels of HAI identification. This is the third paper published since commencing the PhD, and I’m aiming to do at least two, possibly three more before i finish…having said that, it is getting to the pointy end  with now less than 12 months to go (probably 8 months to final seminar, 10 months to thesis examination). Nevertheless, to get there I need to keep ticking off the short term goals, and there’s still plenty of those ahead.

Starting with the write up of my Discrete Choice Experiment. Discrete choice experiments (DCE) are a quantitative attribute based survey method, used to elicit preferences for healthcare products, interventions, services, policies or programs. My application of them in the infection prevention and control setting is novel. Essentially it involved presenting participants two hypothetical surveillance programs and asking them to determine which one they would prefer i.e believe to be most beneficial to their infection prevention program. Sounds simple enough, and really the administration of the survey was probably the easiest part. The hard part of a DCE is deciding what factors will make up your choices. I reviewed the literature and conducted semi structured interviews with international and Australian leaders in surveillance programs. From the data, I identified 5 key attributes of a HAI surveillance program. Then under the supervision of experts from Flinders University in SA, I constructed two blocks of 13 choice questions that would make up the DCE. Together with some attitudinal questions and some basic demographic data, it was all delicately placed into an online survey tool. I purposively selected and contacted leaders in infection prevention and control in Australia and asked them to complete the DCE. The survey was open for five weeks. The high response rate has provided me with confidence that results are meaningful. In early August I spent another few days with my colleagues at Flinders who advised me on the data analysis. And now, I am writing it all up. This data will also be presented at the ACIPC Conference 2015 in November in Hobart.

In the meantime I continue to explore the rich data I collected in the semi structured interviews with the aid of my colleagues who are much more in tune with qualitative data analysis than I am…if I learn nothing during this PhD, I will at least be better equipped to undertake qualitative research! (I did learn today that apparently the ‘proper’ way to say research is not ‘REsearch’ as we tend to say, but rather ‘reSEARCH‘!)

All the while my default is working on my literature review for my thesis, a never ending – always amending beast, and of course thinking about how I will tie all this work up into one cohesive document…!


I cant finish this update without reference to a recent brush with 15 minutes of fame. Recently our dog, Hugo, was struck by two trains receiving nasty leg wounds, a blinding eye injury, and unknown head injuries. The decision was made to treat conservatively, and after two months he has had the last of his dressings, and is getting around ok with modified front legs and one eye. The local newspaper got wind of the story, and within a 24 hour period (whilst I was away doing my DCE stats) Hugo and the family became famous for a day. Online news, two TV news channels, and even a talk show…..

Online news story here

and TV news pics below!

Screen Shot 2015-08-21 at 12.43.07 Screen Shot 2015-08-21 at 12.48.51 Screen Shot 2015-08-21 at 12.50.38

What if…

What if a national HAI Surveillance program is not worth it for Australia? Literature is glowing with many benefits of a national program, but sometimes I get the feeling that true validation of many programs is glossed over. Yes, validation studies are complex and expensive, which is why there are so few. So then, are we really comfortable forging ahead with national surveillance if we’re not totally sure that data is completely true?

How accurate do we need national HAI data to be? The answer to this is, I guess, it depends on its use. Is it used to compare and penalise poor performing hospitals? Then surely it needs to be close to 100% accurate. Is it used for internal purposes, and tracked over time internally as a quality activity? Then probably, we can forgo some accuracy, so long as the method is consistent – which in itself presents a whole new angle. Do those collecting the data always use the same sources? Are those who apply the HAI criteria in total agreement all the time? If there is only a tiny degree of variation in these factors between two surveillance people in one hospital, imagine the degree of difference between 100 surveillance people from 50 hospitals, and how this might effect national data.

It is exciting that the sense of implementation science is now coming to the fore. We cant just develop bundles that reduce infection rates by significant amounts (as demonstrated in robust studies) and say “there you go, do this”. It is clear we need to also address the implementation process to ensure the said bundle is accepted and implemented correctly to facilitate maximum effect.

So this is for surveillance too. My PhD has taken me down the road of identifying the gaps in HAI surveillance in Australia, some of which is published (here and here, more soon hopefully!). With ethics now complete, I will shortly be exploring what attributes of a surveillance program key stakeholders value most. The results from my discrete choice experiment will be fascinating (to me anyhow), and will have a large impact on my final recommendations…

So what if the key stakeholders don’t value a national program but are quite happy doing what they do now? Well I can tell you that some of what is being one now is useless and a waste of precious infection prevention resources. Why it is being done I don’t know for sure, but I suspect at the very least, there is a box somewhere that gets ticked…nevertheless, I believe much across the land can be improved to provide more meaningful data and better guide both local and national interventions.

Will my research produce evidence based recommendations for a national HAI surveillance program? Well yes, on several points. Part of what they provide will be vital data on what key stakeholders value in a national program, and what they don’t value.  This small but crucial bit of information will calibrate our spinning compass, and also plant the first tiny baby steps to successful implementation – should a national program ever be developed.

plodding…

Easter has flown by without the sense of recreation and relaxing of Easters gone by. Its not all nose to the grindstone stuff, but amongst the many small steps along this journey there is occasionally one that causes you to pause, look down, look behind and observe where you’ve come.

My second paper of the PhD, which is the first from my first study on variation amongst HAI surveillance practices has been published in the American Journal of Infection Control. Whilst the original manuscript submitted was returned with a request to downsize resulting in a Brief Report, the final result is a neat punchy little paper, that carries several important messages.

First, despite the fact we produce national data on SAB,  broad variation exists in surveillance practices between State and Territory, public and private, and large and small facilities. Second, only half of the respondents who undertake HAI surveillance have ever received training in surveillance. Third, approximately half of the respondents undertake surveillance retrospectively, despite best practice recommendations that surveillance should be prospective. Finally, less than half risk adjust surgical site infection rates.

So what does this mean? Does this question the validity of the SAB data? (that coincidentally was published today by the National Health and Safety Authority). Could some of the variation in SAB rates be due to this variation in surveillance? Yes and yes. Though the amount of effort the ACSQHC have put into the SAB surveillance, and the reasonably straight forward definitions, I’m guessing less of the difference in rates would be attributable to the variation in practices. This also tells us that surveillance education is more than just deconstructing definitions. We also need to educate about method, data sources and importantly reporting – nationally, to everyone who does HAI surveillance.

Importantly it also means that as we head towards more national surveillance data, we need to ensure that uniform definitions, methodology are adopted, accompanied with uniform education and regular assessments of those charged with running the surveillance programs at a hospital level. All this to be preceded by ensuring that the surveillance program has clear objectives and purpose.  There are many examples, both locally on a small scale, and overseas, on how this can be achieved.

There is a second paper currently under review on the second part of this study, the results from seven clinical vignettes that were included in the survey. Agreement levels were explored, and possibly factors influencing the agreement levels were also investigated.

Now heading toward finalising my DCE and then administering it of course. This will allow some time to write….that’s the plan anyhow!

 

Less than half?

Glass half empty kind of guy? Probably more so than glass half full. But to think I now have less than half of my intended PhD time to go is rather astonishing…I can see the end, but its way past a mountain of work. In fact I still have the majority of work to do.

The festive season saw me focus on my next paper, currently with AJIC which describes some of the differences I found in my survey of ICP staff in Australia who undertake surveillance. More or less a detailed version of my talk at ACIPC 2014. I submitted that on Xmas eve. Then I turned my attention to my next paper, the results of the clinical vignettes which also was part of the survey. I explore the agreement levels of the vignettes but also look for any association with respondent characteristics which might predict accuracy of vignette. The stats in that paper were heavy, but thanks to the advice of statisticians at QUT I was able to find the right test and apply it! Poisson regression. The results were interesting, and will certainly help inform my final recommendations. That paper is now in its final draft whilst we decide where to submit.

Last week I received notice from Australian Health Review that my first article was now in print and online, and the editor had selected it for open online access. This is a good thing as hopefully it will be read more broadly. Further, AHHA asked me to assist with a press release in case there was some media interest…none yet! This was then followed by a statement and portraits for QUT! All amusing as really, I think my next papers will be much more interesting.

(Some coverage on HealthCanal, and an interview with Nursing Review, but if you dont have log in details, you can hear the interview here)

A seemingly major decision which has been met with mixed response, is my plan to now complete my PhD by publication rather than by monothesis. I have several reasons for this:

  1. I have one paper published, one in review, one about to be submitted and I plan to have written another one, possibly two, by the end of this year, finishing up with another early next year. So definitively five, maybe six, papers from my research.
  2. I want my studies to be widely read. Who reads PhD monothesis apart from the examiners and maybe a few other? I’m a PhD student, and I have only read probably half dozen theses…I figure if I publish as I write, more will read…
  3. I believe my work is relevant now…so needs to be published now, not later after I’ve completed my PhD
  4. It just sort of makes sense given the above.

I certainly don’t believe this will be easier or quicker. Some feedback to me has been that some consider this method is taken up by those who want to finish quickly. I don’t agree. I still need to complete my research, I still need to collect data, analyse, write it up etc…And as far as my thesis goes, I still need to provide a detailed lit review, and summary, plus I need to narrate between publications. Its possible this will be harder for me, but as 4 above, it kind of makes sense.

And so soon I head over to Adelaide for a few days at Flinders Uni with my Discrete Choice Experiment experts to begin the construction of the DCE. I have almost completed my semi (or loosely) structured interviews with surveillance experts to assist in identifying the attributes and levels…and then its down that pleasurable road of ethics!!!

We did manage to sneak away to Lorne on the west cost for a week for a bit of a break..the weather was mixed (below), but a week away is precious.

 

IMG_2823

One day it just rained non stop!

and then the next…

IMG_2832

Whilst in transit…

Sitting at Heathrow for five hours after the loooong haul from Melbourne hosts not the most active or creative mind, so thought its time for a catch up.

First, at long last my first “in PhD” paper has been published online at Australian Health Review. Just a brief and simple whip around of what we can discern is happening with HAI surveillance in Australia. It wont change the world but its nice to have it out there.

Second, the reason I’m in transit is I have a poster of my first study to present at Healthcare Infection Society 2014 Conference in Lyon. It has some of the ‘highlights’ from my first study exploring variation in HAI surveillance practices across Australia. Some really interesting (I think) but perhaps not so surprising results. Needless to say there are some issues to address to get a national surveillance program off the ground. Perhaps the most surprising for me was that only half of the respondents who do surveillance have ever been trained in HAI surveillance…that alone leads to so many questions. The first paper from that study is now in draft stage.

Third, I’ve been invited to present my findings at the ACIPC Conference in Adelaide. So literally the day after i return from Lyon get back on the plane. Honored to have been asked, and privileged to share my study. With so much data to report, it is, as always, challenging to find the best hooks to present.

Fourth, the interviews for my attribute discovery have commenced. Already conducted one with a major HAI surveillance leader from Europe, and catching up and interviewing two more whilst in Lyon.

So its a hectic time, and I’ll be glad to see December, which unbelievably brings me to the brink of 18 months of PhD. Though somehow I sense things wont lighten up…

One year in and …

And so on the milestone birthday of the significant other my one year anniversary clocks over. If there ever was a reason to reflect, then clearly this is it….

So in the last year, my first PhD year, i can place ticks in the checkboxes which demonstrates my required progress. Stage 2, tick, literature review, tick, manuscript for publication, tick, conference presentation, tick, confirmation document, tick, confirmation presentation, tick. What a good student.

But the list of awakenings and achievements that aren’t on the checkbox form is longer, and arguably more significant.

The lightness of not being responsible for anybody else but your selfish self….no staff to manage, no endless committees to report the same information over and over again so they can populate their agenda. The sheer relief of not driving three hours a day just to get to work and back.

The surprising offers of bit and pieces of work that come past. Knowing that life as a student (with a mortgage and three dependants) is not financially comfortable, colleagues seem to nominate me as the go to person to pick up stuff that any full timer doesn’t want to commit to.

The power to direct the work so it can be applied in the real world. The knowledge of a late in life student had of real world politics allow me to understand more precisely what is useful, meaningful and applicable in the real world.

The acceptance that sometimes what is called the real world, is in fact, just the fake world. The real real world is actually much more fun and interesting.

Understanding how much i don’t know, but importantly appreciating what it would require for me to know.

The networking with some very clever people, and the impressiveness of youth and thinking big.

There are many others, and as my Thesis Whisperer points out, I should be embracing all these things more, but for the pressures of ticking boxes…

 IMG_1957  

the cakeIMG_1917

my home office on a lovely sunny morning