1 Respondents’ characteristics
All 18 physicians and nurses we invited agreed to participate (see Table 1). The sample included both users (n=16, organizations A-C) and non-users (n=2, organization D), and covered somatic, psychogeriatric as well as rehabilitation care. The majority of participants were female (14 out of 18). The median number of years of work experience was 17 (range 0.5 – 35 years).
2 Role of POCT in urinary tract infections
2.1 Only added value in case of non-specific symptoms
About half of the respondents explicitly stated that, in nursing home residents presenting with specific urinary symptoms, inflammatory marker POCT (referred to below as POCT) has no added value as there is no diagnostic uncertainty,
but that attaining such diagnostic certainty was only possible in a small minority of nursing home residents able to adequately express complaints.
Well, if you’re not in doubt, then there’s no need, is there? If someone shows the full set [of symptoms], then in my view there’s no point in confirming it with POCT. A-IV
Then it’s fairly obvious. But the group who are able to express this clearly and accurately is very small. B-IV
The other respondents implicitly agreed with this limitation of the scope of POCT, as they only discussed testing in the context of residents with non-specific symptoms.
2.2 Objective measure for ‘ruling out’ UTI
All respondents primarily expected POCT to decrease diagnostic uncertainty by ‘ruling out’ UTI in residents with non-specific symptoms. Negative test results were considered to be an objective measure of the absence of UTI.
But sometimes you’re faced with people where you think: well, this is really very very different from what we usually see. And you then want to exclude [a UTI] some way or other, as it might still be that. A-II
So you want to have something more objective, as that history taking [for UTI] is so problematic, something a bit more objective regarding the severity of illness, to help you decide. So that [if the POCT] is below 20, you think, well, that you, that we can just as well wait and see. A-VIII
But several respondents expressed concerns about false negative test results, which could lead to erroneously withholding antibiotic treatment in nursing home residents with a UTI.
So I test it, and it’s a CRP [POCT] of 2. So at any rate you know it’s not an obvious infection. [But] I’m not sure it’s that reliable. A-V
The danger might be that someone is not treated who should be treated, as they would otherwise become more severely ill of their complaints persist. C-I
Having POCT as an objective measure in addition to clinical reasoning was perceived as important to convince others, such as family members and professional caregivers, of the appropriateness of withholding antibiotic treatment from residents with non-specific symptoms.
And you think: you know what, let’s do a POCT, and if it’s zero, you can then say well, this is definitely not it. So you have a stronger argument. […] You have an objective finding, in addition to what I see or hear myself. Not that I think that’s not reliable, but relatives often want a test or a scan. B-II
However, some respondents also admitted that POCT as an objective measure was of additional importance to corroborate their own clinical reasoning, for instance because it could support them in taking another direction in their diagnostic reasoning than just considering UTI.
You want to check that your clinical impression is right. We doctors like that. To have some proof. A-VI
And that enables you to change direction sooner with this, well, in this area. A-II
Although POCT was perceived as helpful for ‘ruling out’ UTI in nursing home residents with non-specific symptoms, their symptoms generally persisted. Respondents reported that in such cases diagnostic uncertainty remained. Hence, this could induce a feeling of unease, which in turn could lead to prescribing antibiotics after all.
But I would be […] skeptical at first. Like, well, the CRP [POCT] is lower, so now I can’t treat them for UTI. [But] what then?C-I
And so you either start to treat because you feel you have to do something […] Or you don’t, and then the complaints often persist anyway… And so at a later stage you still start to treat. A-V
2.3 Objective measure for ‘ruling in’ UTI
Almost all respondents additionally thought POCT was a useful tool for ‘ruling in’ UTI. The inability of many nursing home residents to adequately express their symptoms made it difficult to rely on clinical reasoning alone. As an objective measure, POCT could reduce this diagnostic uncertainty, since positive test results would confirm UTI diagnosis and justify antibiotic treatment.
That would be really helpful, as this remains somewhat fuzzy, especially in our target group. […] You can do an abdominal exam. But is it tender: is it constipation or a bladder infection? Or do they just dislike the fact that you’re touching their belly? So for us, diagnostics is still a kind of intuition, and such a test would actually… A-IV
And if I then also find an elevated CRP [POCT], that would be an argument to say this is really a UTI. As I don’t have any other instrument.B-I
Some respondents also thought that POCT could speed up the diagnostic process, and thus the start of antibiotic treatment for UTI, as it is a ‘rapid test’, and its objective nature could compensate for suboptimal clinical attention.
Perhaps it’s more a matter of seeing that if the POCT is elevated, you’ll give an antibiotic sooner, whereas that person would only become really ill tomorrow; so you […] can get a timely start. D-I
But at [psychogeriatric wards] for instance, where there is less close clinical surveillance, there might be a group that you could treat a bit sooner. A-V
Yet, in order to fulfil its supporting role in ‘ruling in’ UTI, it was acknowledged that POCT values should exceed a certain threshold in residents with a UTI, even in residents with non-severe illness or presenting with non-specific symptoms. The latter was doubted by some respondents.
People may not be completely healthy, but not terribly ill either, so you think: it might be a UTI. But whether [ a POCT] would really rise that much in that kind of infection […] I wonder. B-IV
Despite the perceived advantage of POCT for ‘ruling in’ UTI, various respondents thought that positive test results might give rise to new diagnostic uncertainties. Due to the limited specificity of general inflammatory markers, positive results could also indicate the presence of other infections. Therefore, treatment decisions remained dependent on clinical reasoning.
All these things indicate that there may be an infection present. […] You don’t know where the infection is located. It doesn’t have a specific … it could be anywhere. […] It could be the airways or other types of infection. Skin infections are very common. You get all kinds of infections with the same signs. […] You have to look at that carefully. […] So [POCT] is just an addition to the toolkit. B-I
As a result, some respondents feared that the limited specificity of a POCT could lead to inadequate treatment, such as prescribing antibiotics for UTI while the resident is actually suffering from another infection.
If you have someone with vague symptoms and a high [POCT], do you then know it’s a urinary tract infection? […] The danger is that there’s an infection elsewhere, and that you treat them with nitrofurantoin and that that has no effect. B-IV
Hence, in the case of non-specific symptoms, some respondents recommended either combining POCT with further examination of clinical symptoms to exclude infections other than UTI, or not to use POCT at all in residents suspected of having another infection besides UTI.
You preferably want to get it in context as much as possible. So not someone who also happens to have a cough and has just arrived on the rehabilitation ward, and might have a cystitis. Cos then you don’t know what you’re measuring. A-V
2.4 Balancing POCT and clinical reasoning
POCT was thought to reduce diagnostic uncertainty if the test results matched the respondents’ clinical reasoning. In case of disagreement between the two assessments, respondents reported different ways of balancing them against each other. Sometimes, clinical reasoning prevailed: respondents explained how they would either refrain from antibiotic treatment in residents who were not very ill despite positive tests results, or would start such treatment in severely ill residents despite negative test results.
When I see a CRP of 100 and the patient doesn’t appear ill at all, I will not refer them to hospital straight away purely on [the basis of] this CRP. A-V
Sometimes you might do a POCT…And then you think, well, that’s not very high. And still the patient is so ill that you give them an antibiotic anyway.D-II
In other situations of disagreement, POCT results would overrule clinical reasoning. For instance, some respondents reported they would start antibiotic treatment in response to positive test results in residents who were not very ill. An important reason to do so was the general frailty of the resident.
And in case of doubt; if I do see an elevated CRP [POCT] that makes me think right, there’s something going on there. And especially if it’s a very vulnerable patient, then I’ll think: let’s start treatment anyway. B-II
When balancing the results of POCT with the clinical work-up, some respondents acknowledged the importance of support for nursing home staff in the interpretation of POCT, given the characteristics of a particular test.
I think it’s more important for us to get clinical training about how to interpret test results and when to use them. […] How about the specificity and sensitivity, can you use it to rule out, can you use it to rule in? B-III
You ought to know the predictive value of a test. […] It’s just that we tend to forget that all the time. D-II
Several indications other than diagnosing UTI were mentioned for POCT. Firstly, respondents suggested that POCT could be useful to (1) demonstrate if a resident is suffering from any infection in general, as well as to (2) confirm specific infections that are common in nursing home residents.
And also to decide whether someone is not the way they usually are, or we can’t find out exactly what’s the matter. Then it’s sometimes interesting to know, well, could [it] have to do with an infection? A-VIII
That would be good, as for instance cellulitis, that’s sometimes an unclear picture here at the nursing home. […] Then you have a much more powerful argument to start antibiotics. A-V
Secondly, respondents explained that POCT could be helpful in distinguishing infectious from non-infectious disorders with comparable clinical manifestations.
Of course sometimes you get vague abdominal symptoms that make you think: is this a cholecystitis or gall stones or what? I could imagine that you might be able to differentiate with that too. A-IV
Thirdly, POCT was regarded as a useful tool to indicate infection severity, to monitor the course of an infection, or to predict a resident’s prognosis in general. In such cases, POCT could, for instance, help to decide whether to consult a specialist or whether to refer a resident to an emergency department.
Yeah, sometimes you get people who deteriorate, and you can’t put your finger on it. Is it an airways infection rather than a urinary infection? So you’re at a loss. And you then check which way it’s developing. For prognosis too. A-IV
If people are ill and you ask yourself can I safely leave this patient here or should I refer them to an orthopedist for a check-up as we’re thinking it might be an infection of the hip, for instance. D-II
At the same time, most respondents questioned the evidence base for these other roles, and noted that cut-off values for interpreting POCT results should be, but were currently not, available for these additional applications.
Except, well, it’s not really intended for, we don’t have official cut-off values. Well. So then it’s still your own interpretation of the overall… A-VIII
Well, so where’s the threshold? I think it’s very important to put it in context and link it to the evidence base. A-V