Using Point-of-care C-reactive protein to guide Antibiotic prescribing for Respiratory tract infections in Elderly nursing home residents

Antibiotics are among the most commonly prescribed drugs in nursing homes
(NHs), with 47% to 79% of the residents receiving at least one antibiotic
course annually, while almost 25% of these prescriptions do not seem required
at the index consultation. Inappropriate antibiotic use is undesirable both on
the patient level (because of exposure to side-effects and drug interactions)
and the societal level (because it contributes to the development of antibiotic
resistance).

About one-third of the antibiotic prescriptions in Dutch NHs are for treatment
of suspected lower respiratory tract infections (LRTl). Diagnosing LRTl is
challenging in this setting as NH residents often lack typical symptoms, and
cognitive disabilities can impede communication of experienced complaints. ln
addition, X-ray, bacterial culture and laboratory research are often not
well-applicable in the vulnerable NH population or not directly available in
this setting. The resulting diagnostic uncertainty leads to AB prescribing to
be 'better safe than sorry', an attitude towards antibiotic prescribing that
can be further enforced by a variety of external factors, such as expectations
of patients, family members and nursing staff. Together, these factors increase
the chance of unnecessary antibiotic prescribing for NH residents.

ln general practice, diagnostic uncertainty and patient expectations play a
role as well in treatment decisions regarding LRTl. ln this setting, it has
been investigated whether the use of C-reactive protein (CRP) *point-of-care
testing (POCT) can improve antibiotic prescribing for LRTl. With CRP POCT,
capillary blood tests are performed with test results available within minutes:
the resulting CRP values are an indicator of presence and severity of
inflammation. Several studies demonstrated that adding CRP values to the
evaluation of clinical signs and symptoms indeed improves the reliability of
the diagnosis pneumonia. The use of CRP POCT in general practice showed a
reduction in antibiotic use in patients with suspected LRTI. This appeared
cost-effective in comparison with usual care. These findings were confirmed by
other recent studies, which resulted in the incorporation of the CRP POCT in
the guideline 'Acute cough' of the Dutch College of General Practitioners
(NHG), and consequently in the large-scale implementation of CRP POCT in
general practices in the Netherlands.

CRP POCT may be a promising tool to reduce unnecessary antibiotic use for LRTl
in NHs as well, considering the large degree of diagnostic uncertainty and
other factors that influence antibiotic prescribing for this population.
Previous studies indeed showed that CRP is an useful indicator for the severity
of LRTI in NHs, However, the effectiveness of CRP POCT on antibiotic
prescribing has not been investigated in NHs. This is the main reason for the
current limited use of the CRP POCT in NHs, despite its incorporation into the
NHG guideline, which is also acted upon by physicians in NHs.

We hypothesize that the implementation of CRP POCT for suspected LRTl in NHs
will lead to a clinically relevant reduction in antibiotic use of at least 15%,
without any negative consequences for patient recovery.

CENTRAL OBJECTIVE
To investigate whether the use of CRP POCT results in a reduction in antibiotic
prescribing for NH residents with suspected LRTl, without any negative
consequences for patient recovery.

SECONDARY OBJECTIVES
1) To explore potential associations between CRP POCT values and A) signs and
symptoms in NH patients with suspected LRTl, B) antibiotic treatment.
2) To investigate cost-effectiveness and cost-benefit of the use of CRP POCT in
the NH setting.
3) To investigate potential facilitators and barriers for the use and
implementation of CRP POCT in the NH setting.

CENTRAL RESEARCH QUESTION
A cluster randomized controlled trial (cRCT) will be conducted in a 1.5-year
period to answer the central research question. 10-12 nursing home (NH)
organizations will be recruited to participate in the cRCT. Allocation of
control and intervention groups will occur with randomization at NH
organization (cluster) level.

Participating NH organizations will be randomized to an intervention or control
group. In both groups, care as usual is provided, however, in different forms.
ln the control group, suspected LRTI are evaluated based on clinical signs and
symptoms, possibly supplemented with additional diagnostics such as a CRP
laboratory test using blood collected by venipuncture, culture or X-ray (the
latter two being only sporadically applied in nursing homes). ln the
intervention group, physicians can use CRP POCT * a different, less commonly
used form of usual care, in addition to their evaluation of clinical signs and
symptoms and possibly other additional diagnostics. There are three repeated
measures where case report forms are completed en submitted by physicians
within the electronic patient file: at the index consult (where LRTI is
diagnosed, with or without the aid of CRP POCT), and two follow-up measurements
at 1 and 3 weeks after the index consultation. In addition, pharmacy data for
each NH will be collected retrospectively on total antibiotic use in the NH
during the study period.

SUB RESEARCH QUESTION 1
Data of patients in the intervention group of the cRCT will be used to
investigate possible associations between CRP POCT values and A) signs and
symptoms in NH patients with suspected LRTl (i.e. do CRP POCT values correlate
with certain signs and symptoms?), and B) antibiotic treatment (i.e. what are
the ranges of CRP POCT values within which physicians decide to prescribe
antibiotics?).

SUB RESEARCH QUESTION 2
An economic evaluation will be performed from a societal perspective, using
cost-effectiveness and cost-benefit analyses.

SUB RESEARCH QUESTION 3
We will conduct a facilitator and barrier analysis on the use and
implementation of CRP POCT in the NH setting. Semi-structured qualitative
interviews and focus groups will be conducted with both NH professionals (e.g.
physicians, nursing staff) and NH management. A topic list will be developed,
with three dimensions of factors that influence the use of CRP POCT: on the
cultural (e.g. financial mechanisms), contextual (e.g. organizational factors)
and the behavioral (e.g. attitudes) level. Also included in the topic list are
potential barriers and facilitators for the use of CRP POCT specifically for
mentally incompetent nursing home residents.

Data collection will continue until data saturation has been achieved (i.e. no
new barriers and facilitators are being identified), and to this end data
analysis will coincide with data collection.

ln the control group, care as usual is provided for patients with suspected
LRTI: the evaluation of suspected LRTI is based on the examination of clinical
signs and symptoms, possibly supplemented with additional diagnostics such as
CRP laboratory testing using blood collected by venipuncture, culture or X-ray
(the latter two only being sporadically applied in nursing homes). In the
intervention group, physicians have access to CRP POCT, and evaluate suspected
LRTI based on clinical signs and symptoms combined with CRP POCT results and
possibly additional diagnostics. The CRP POCT is performed by measuring the
concentration CRP in a finger prick blood, which indicates the presence and
severity of inflammation.

The reference value of CRP is <10mg/L for healthy persons. Reference values in
elderly people with comorbidity are marginally deviant from the reference value
in healthy people, with on average only a couple unit increase in mg/L due to
comorbid conditions. The CRP-value increases rapidly (within 6-8 hours) after
inflammatory stimuli, and a peak level is reached after approximately 48 hours.
The guideline "acuut hoesten" of the Dutch Society of General Practice*
includes the following information on cut-off values of CRP in cases of
suspected LRTI: a CRP-value of <20 mg/L indicates that pneumonia can be
excluded with great certainty, and a CRP >100 mg/L strongly suggests pneumonia.
In case of a CRP-value of 20-100 mg/L the presence of certain clinical signs
and symptoms, and other risk factors, have more weight in the treatment
decision-making.

Prior to the study onset, physicians from NH organizations in the intervention
group will receive a training on the background and use of CRP POCT, and
interpretation of CRP POCT results. Part of this training will involve
information on ways to stimulate the blood circulation in the patient's
fingers, as decreased blood circulation may hamper the procurement of a blood
sample. The following methods to stimulate blood circulation, with due
consideration to the patient's state of mobility, will be discussed: asking the
patient to wave his/her arms, clap his/her hands, wash his/her hands in warm
water, or hold a glove with warm water. If available at the site of the NH, an
infrared lamp can be used to warm the patient's hand.

* With reference to information on CRP POCT in the guideline "acuut hoesten" of
the Dutch Society of General Practice ("Nederlands Huisarts Genootschap"):
https://www.nhg.org/standaarden/volledig/nhg-standaard-acuut-hoesten

Nursing home residents participating in this study will experience minimal
burden. The number and timing of physician consultations are not impacted
negatively by participation in the study. Also, patients are not involved in
the data collection themselves (this occurs through the electronic patient
file). In the intervention group, patients will undergo blood collection via
finger prick once at the index consultation and if necessary at follow-up
consultations. The burden the patient will experience from this finger prick is
inconvenient at most, given that other types of diagnostics using finger prick
blood collection are common practice in nursing homes (D-dimer, blood glucose)
without any problems. The finger prick will be discontinued for patients who
hitherto consented to participate, but at the consultation show signs of
resistance towards the finger prick. The physician judges whether resistance
(verbal and/or non-verbal behavior that is a more excessive manifestation of
the usual reaction to situations deviant to daily routine) is the case in the
instance, and consults with other caregivers if necessary. In case of
resistance, the patient will receive usual care without the CRP POCT (using
finger prick blood collection).

Participants do not have risk of (minor) harm/injury because of blood
collection via finger prick. A possible scenario where the participant may have
risk of an inappropriate treatment policy and consequently higher
morbidity/mortality risk involves a prematurely timed CRP POCT. This situation
would entail the CRP POCT being performed at a time where CRP values are still
rising, resulting in non-prescribing or delayed prescription. However, this
risk is minimal, which is also apparent from the application of CRP POCT in
general practice, given that physicians do not solely rely on CRP-values in
their decision on treatment policy but rather use CRP POCT as an additional
tool in a spectrum of clinical findings. If patients show signs of severe
illness, they are always followed-up by health care professionals/ the
physicians, even in the case of low CRP-values at earliest consultations.
Consequently, changes in clinical signs/symptoms can be adequately followed-up
using additional diagnostics and/or changes in treatment policy. These
scenarios will be given its due consideration in the training physicians from
participating (intervention group) nursing homes will receive to adequately
apply and interpret CRP POCT results.

If the hypothesis of the UPCARE study turns out to be correct, LRTI patients in
the intervention group will receive improved care (adequate and timely
treatment) because of improved diagnostics. Improved care entails less
unnecessary antibiotic prescriptions and thus less side effects, medicine
interactions, and antibiotic resistance. Also, if patients have different
underlying diseases (other than LRTI/pneumonia, not requiring antibiotic
treatment) they will receive treatment more appropriate to their signs and
symptoms. Also, the CRP POCT can be considered a less invasive intervention
compared to the CRP laboratory test, with respect to the type of blood
collection (via finger prick respectively venipuncture). Study participation
does not encompass specific advantages for patients in the control group.