JAMDA 23 (2022) 968e975

JAMDA
journal homepage: www.jamda.com

Original Study

Implementation and Use of Point-of-Care C-Reactive Protein Testing

in Nursing Homes
Tjarda M. Boere MSc a, Rogier M. Hopstaken MD, PhD b,
Maurits W. van Tulder PhD c, Franҫois G. Schellevis MD, PhD d, e,
Theo J.M. Verheij MD, PhD f, g, Cees M.P.M. Hertogh MD, PhD a, g,
Laura W. van Buul PhD a, *
a
Department of Medicine for Older People, Amsterdam Public Health Research Institute, Amsterdam UMC, VU University Medical Center, Amsterdam,
the Netherlands
b
Star-shl Diagnostic Centers, Etten-Leur, the Netherlands
c
Faculty of Behavioral and Movement Sciences, Amsterdam Movement Sciences Research Institute, Vrije Universiteit Amsterdam, Amsterdam, the
Netherlands
d
Department of General Practice, Amsterdam Public Health Research Institute, Amsterdam UMC, VU University Medical Center, Amsterdam, the
Netherlands
e
NIVEL (Netherlands Institute for Health Services Research (NIVEL), Utrecht, the Netherlands
f
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands
g
National Institute for Public Health and the Environment (RIVM), Bilthoven, the Netherlands

a b s t r a c t

Keywords: Objectives: This study evaluated logistics, process data, and barriers/facilitators for the implementation
Process evaluation and use of C-reactive protein point-of-care testing (CRP POCT) for suspected lower respiratory tract
antibiotic stewardship infections (LRTIs) in nursing home (NH) residents.
C-reactive protein point-of-care testing
Design: This process evaluation was performed alongside a cluster randomized, controlled trial (UPCARE
lower respiratory tract infection
study) to evaluate the effect of CRP POCT on antibiotic prescribing for suspected LRTIs in NH residents.
nursing home
Setting and Participants: Eleven NHs in the Netherlands.
Methods: Data sources for process data regarding intervention quality included a questionnaire among
NH staff, logs, reports, and CRP POCT-analyzer records. Barriers and facilitators for implementation were
assessed in focus group interviews with physicians and nurses from 3 NHs.
Results: Correct patient selection for CRP POCT and generally continued CRP POCT use indicated good
fidelity. The initial training and training of new employees seemed to fit the need, but some POCT-user
group sizes had increased over time, which could have impeded frequent use. Users were generally
satisfied with CRP POCT and perceived its use feasible and relevant. Facilitators for implementation were
initial commitment and active initiation, followed by continued attention and enthusiasm for building
routine practice and trust. Short lines of communication between staff, short distance to the POCT-
analyzer, 24/7 coverage of staff, and a clear task division facilitated continued attention and routine
practice.
Conclusions and Implications: This process evaluation showed sufficient quality of providing CRP POCT in
Dutch NHs. We processed findings of intervention quality and implementation knowledge into key
recommendations for CRP POCT implementation in this setting. Future research could focus on CRP POCT
use in countries with different organization of care in NHs.

This work was supported The Netherlands Organization for Health Research and Development (ZonMw, program Antibiotic Resistance, grant number 541001 004).
LB reports a grant from The Netherlands Organization for Health Research and Development (ZonMw, program Antibiotic Resistance, grant number 541001 004), for the
conduct of the current study. TB, FS, CH, RH and MT have no competing interests to disclose. TV reports grants from Abbott, Becton Dickinson, Biomerieux, European
Commission, grants from European Commission, Orion, during the conduct of the study; grants from European Commission, Janssen Pharmaceuticals, grants from ZonMw,
outside the submitted work. The Medical Ethical Committee of the VU University medical center in Amsterdam approved the trial protocol on March 28, 2018, as well as
participation of all recruitment sites (NHs) in the trial. The trial was registered at the Netherlands Trial Register on 29 August 2018 with trial number NL5054. Orion
Diagnostica and Saltro made in total 20 CRP POCT instruments available for the study.
* Address correspondence to Laura W. van Buul, PhD, Department of Medicine for Older People, Amsterdam Public Health Research Institute, Amsterdam UMC, VU
University Medical Center, Room MF C-370, Van der Boechorststraat 7, 1081 BT Amsterdam, the Netherlands.
E-mail address: l.vanbuul@amsterdamumc.nl (L.W. van Buul).

https://doi.org/10.1016/j.jamda.2021.09.010
1525-8610/Ó 2021 The Authors. Published by Elsevier Inc. on behalf of AMDA e The Society for Post-Acute and Long-Term Care Medicine. This is an open access article under
the CC BY license (http://creativecommons.org/licenses/by/4.0/).
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 969

Ó 2021 The Authors. Published by Elsevier Inc. on behalf of AMDA e The Society for Post-Acute and
Long-Term Care Medicine. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).

Common infections in nursing home (NH) residents such as lower secondary outcome measures involved clinical outcomes during
respiratory tract infections (LRTIs) are difficult to diagnose due to follow-up.
atypical clinical presentation in the older patients and limited diag-
nostic options in this setting.1e5 Given the potential severity of clinical Intervention Description
outcome, antibiotics are often over-prescribed for LRTI to be “better
safe than sorry.”5e9 Point-of-care testing (POCT) to rapidly assess the In Table 1 we summarize the various elements of the CRP POCT
level of acute inflammatory marker C-reactive protein (CRP) might process and corresponding intervention components. Our study pro-
improve appropriateness of antibiotic prescribing by providing phy- tocol article20 provides a more detailed description of intervention
sicians information on presence and severity of infection.8,10e15 components, such as medical and technical POCT training sessions.
POCT can be defined as the process of indicating, executing, pro- The medical training session aimed to provide intervention group
cessing, interpreting, documenting, reporting, and following up of a physicians with sufficient knowledge on correct utilization of CRP
laboratory test by a skilled and competent health care professional in POCT, based on literature and the current Dutch LRTI NH guideline.26
proximity of the patient.16 In primary care, CRP POCT is a proven cost- The 2 main aspects of this training were patient selection for CRP POCT
effective tool for reducing antibiotic prescribing for patients with (ie, in case of clinical ambiguity and potential for therapeutic conse-
suspected LRTIs, including acute exacerbations of chronic obstructive quences) and interpretation of results (eg, cut-off CRP-values for
pulmonary disease. However, evidence for its (cost)-effectiveness in management decisions). Throughout the trial, newly employed phy-
the NH setting is yet lacking.12,17e19 We examined the effect of CRP sicians received the medical training information via our instruction
POCT on antibiotic prescribing for NH residents with suspected LRTI in materials, additional training sessions, or from colleagues. The labo-
a cluster randomized, controlled trial (the UPCARE study: “Using ratory provided a technical training session for potential users of CRP
Point-of-care C-reactive protein to guide Antibiotic prescribing for POCT. For CRP POCT-user group selection (nurses and/or physicians),
Respiratory tract infections in Elderly nursing home residents”). A full we asked NHs to balance full shift coverage (ie, including out-of-hours
description of the UPCARE study design is given elsewhere.20 shifts) and minimal group size to ensure frequent POCT utilization
The implementation and use of a new intervention in NHs can be over time. Among training participants, an in-house POCT-trainer was
challenging, given the heterogeneity of the population and the appointed for the NH to expand the POCT-user group if necessary. The
complexity of the setting.21e23 The implementation process of CRP instructors provided instruction manuals to be kept next to the POCT-
POCT in the UPCARE study required a collaborative approach between analyzer.
the NH, research team, and a primary care laboratory (hereafter The POCT-analyzer was placed in or near the office of the POCT-
referred to as laboratory), to tailor implementation aspects to orga- user group, or in another type of staff room that allowed the POCT-
nizational practices. A process evaluation of this intervention may analyzer to be near the fridge-stored test materials. In general, the
help to obtain insight into the internal and external validity of the blood sample was collected at the resident’s room by the POCT-user.
intervention. Also, findings help to increase our understanding of the Then, the sample was dispensed in the buffer solution in the
magnitude and direction of effect of the intervention and its cuvette, which was then closed off with a reagent cap and transported
implications.21,23e25 to the analyzer for further processing. The test could be performed
The aim of this process evaluation is (1) to determine intervention when the cuvette was warmed up to room temperature and within a
quality regarding CRP POCT utilization in the UPCARE study, and (2) to maximum of 2 hours. The POCT-analyzer Quikread go (Aidian Oy,
determine barriers and facilitators for the implementation and utili- Espoo, Finland) was shown to have adequate analytical performance
zation of CRP POCT in the NH setting. compared with laboratory assessment.27e29 The laboratory is
compliant with the ISO Norm 15189 (M037) and recommendations for
Methods POCT, and verified each POCT-analyzer before transport to the NH site.
The laboratory also continued quality assurance with periodic quality
This mixed-methods process evaluation is part of the UPCARE control visits of the POCT-analyzer at each site.
study and was conducted around the end stage of the UPCARE trial (ie,
March through July 2020). Process Evaluation

Trial Design We evaluated process data related to intervention quality24,25 (ie,

logistics and experiences regarding CRP POCT in the UPCARE trial). For
The process evaluation was conducted alongside a cluster ran- insight into implementation knowledge, we assessed barriers and
domized, controlled trial. This trial was performed in 11 NHs across facilitators for the implementation and utilization of CRP POCT.
the Netherlands, of which 6 were randomized to the intervention
group and 5 to the control group. The study population comprised NH Intervention Quality
residents of psychogeriatric, somatic, geriatric rehabilitation, and
short-term residential care wards, who were suspected of having We assessed intervention quality by evaluating experiences of
LRTIs. Physicians in the intervention group had access to CRP POCT for involved NH employees regarding CRP POCT satisfaction, relevance,
NH residents they suspected of having LRTIs. The control group was and feasibility in the NH setting during the trial. Another parameter of
provided usual care without the use of CRP POCT. Physicians filled out intervention quality was fidelity, which concerns the degree to which
information for participants on clinical status, diagnostic work-up, the intervention was delivered as intended.24,25
and management in electronic case report forms at 3 time points: at Besides process description, data sources were the following: trial
initial consultation, and 1 week and 3 weeks later. The primary data (on illness severity), trial monitoring reports, logs of communi-
outcome measure was antibiotic prescribing at initial consultation and cation with NHs and the laboratory, POCT-analyzer records (last 100
970 T.M. Boere et al. / JAMDA 23 (2022) 968e975

Table 1
CRP POCT Process and Corresponding Intervention Components

CRP POCT Process Component Description Corresponding Intervention Components

Patient selection Appropriate patient selection for CRP POCT by treating Medical training session
physician
Execution A skilled physician/nurse obtains a finger prick blood sample Technical POCT training session; logistics and quality assurance.
and handles further sample preparations. Materials are stored
as per guidelines.
Processing The skilled physician/nurse performs the measurement with Technical POCT training session; logistics and quality assurance.
the POCT instrument within 2 h after blood sampling.
Documentation If different NH employees involved: the skilled physician/nurse Logistics and quality assurance
who performed the measurement communicates the CRP-
value to the treating physician. The treating physician
documents the CRP-value in the electronic health record.
Interpretation The treating physician interprets the CRP-value alongside Medical training session
clinical signs and symptoms and decides on (antibiotic)
management accordingly.
Reporting The treating physician communicates management decisions to N/A
relevant parties involved as per usual care.
Follow-up The treating physician monitors clinical status and uses CRP Medical training session
POCT for follow-up if necessary.

measurements per POCT-analyzer), and a process evaluation ques- Audio recordings were transcribed verbatim. Data analysis con-
tionnaire completed by physicians and nurses of intervention NHs sisted of open coding performed by 2 researchers independently (T.B.
(hereafter “questionnaire”). and M.E.), and axial and selective coding performed by 1 researcher
POCT-analyzer records were retrieved within 1 month after trial and reviewed by 2 other researchers (T.B., F.S., and L.B.). The software
conclusion and comprised pseudonymized user level information (ie, program Atlas.ti v 8 was used. During open coding, the researchers
users filled out their initials). An estimation of the number of users per (T.B., F.S., and L.B.) coded relevant fragments of each transcript indi-
organization and location was made based on initials. POCT-analyzer vidually and discussed codes after each transcript for consensus.
records were used to show the use of CRP POCT over time. During axial coding hierarchical connections were made between
The questionnaire comprised questions on CRP POCT utilization categories through discussion between the researchers (T.B., F.S., and
and appraisal. Most of the questions were close-ended and included a L.B.) about the open codes. During selective coding, data were further
score option between 1 (totally disagree) and 10 (totally agree). structured and researchers reflected on relations between themes and
Questions on satisfaction and relevance included the medical and categories. If relevant, themes and categories were presented in a
technical POCT training sessions, experienced capability of using and conceptual framework.
interpreting CRP POCT, the added value and influence of CRP POCT on
management decisions, and user-friendliness of CRP POCT. Questions
regarding feasibility concerned the use of guideline-specified cut-off
Results
CRP-values, accessibility of CRP POCT (materials), and experienced
feasibility of CRP POCT within the timeframe of initial consultation
Intervention Quality
and during nights and weekends. Questions concerning fidelity
included training attendance, patient selection, and the utilization of
Fidelity
CRP POCT.
Rate of staff involvement. The average attendance rate of all phy-
Descriptive statistics of POCT-analyzer records and the question-
sicians per NH organization to the medical training session prior to
naire were obtained using Microsoft Excel 2016 and SPSS statistical
trial commencement was 39% (range 15%e92%). An optional quiz
software v 26 (IBM Corp).
regarding the medical training session was filled out on average by
32% (10%e60%) of all physicians per organization, irrespective of
training attendance. The average quiz score per organization was
Implementation Knowledge 66% (40%e92%) correct. During the trial, 4 of the 6 intervention NHs
received an extra medical training session 1 or 2 times for new
Focus group meetings took place in June and July 2020 with phy- employees or as a refresher. Two NHs arranged for training of new
sicians and nurses of 3 intervention NHs. A purposive sampling physicians themselves using our instructions. Among questionnaire
method was used to select organizations based on variation in in- respondents, 14% did not receive the medical training information
clusion rate during the trial (low, medium, and high), ward sizes in any form.
(psychogeriatric, somatic, and geriatric rehabilitation), and physician On average 7 (range 4e11) NH employees participated in the
turnover (high-low). The meetings were held using a semistructured technical CRP POCT training sessions. Each organization appointed
topic list and were led by a moderator. All focus group participants 2e4 employees as in-house POCT-trainer. According to available data,
were asked for informed consent and approval of audio recording of in-house POCT-trainers from 2 NHs trained additional (11 respectively
the focus group by the researchers. 13) employees shortly after the training. The run-in period was on
The topic list was developed by the researchers and comprised 2 average 7 weeks (range 2e12 weeks). At the end of the run-in period,
domains: implementation and utilization. Implementation topics the ratio of POCT-users to the number of residents per location (or
included experiences with the medical and technical POCT training organization, as appropriate) was on average 1:26 (range 1:4e1:61).
sessions and organizational culture regarding innovation. Utilization We do not have information regarding newly trained employees
topics regarded experiences with CRP POCT processes and patient during the trial. Of the 21 questionnaire respondents who were
selection, and the added value and influence of CRP POCT in man- trained for CRP POCT, 18 received the training from the laboratory and
agement decisions. 3 from a colleague.
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 971

POCT-analyzer records showed on average 27 users per organiza- use a new sample depending on the type of error (n ¼ 3), or to repeat
tion (range 15e45), with an average of 5 (range 2e12) measurements the test (ie, unspecified for resampling or retesting the sample, n ¼ 2).
per user. Employees who were trained before the start of the trial
performed 48% (range 20%e79%) of all measurements per organiza- Satisfaction, relevance, and feasibility
tion. The proportion of users who performed
5 measurements per Supplementary Table 1 shows the mean agreement scores of re-
organization was on average 71% (range 37%e89%). spondents regarding statements of CRP POCT relevance, feasibility,
and satisfaction. Respondents scored on average >7 regarding ques-
Patient selection. Among the 25 questionnaire respondents who tions on satisfaction with training sessions and user-friendliness of the
used CRP POCT, 5 respondents used CRP POCT for “not ill” patients analyzer. Respondents who received the medical training from a
(according to subjective clinical judgement30), all used CRP POCT colleague or via instruction materials in absence of a live session were
for patients who were moderately ill, and 11 respondents used CRP less satisfied with the method (score: 6.6) and content (score: 6.4).
POCT for patients who were severely ill. This finding corresponded Each agreement score was on average >8 regarding questions of
with CRP POCT utilization primarily for trial participants who were added value and trust regarding CRP POCT, and use and usefulness of
moderately ill: 78% of CRP POC-tested participants were moder- cut-off values for management decisions. Supplementary Table 3
ately ill, 16% were not ill, and 7% were severely ill. shows that >20 of 25 respondents believed CRP POCT utilization to
be feasible at initial consultation, weekends, and nights. Of the 6 NH
CRP POCT utilization. According to the analyzer records, CRP POCT organizations in the intervention group, all wished to continue CRP
was on average across NHs most often used during office hours POCT in all or some of their locations, and 1 immediately continued
(87%, range 68%e96%) and less frequently during nights (7%, range CRP POCT in 1 of their locations after the trial had ended.
1%e15%) or weekends (6%, range 2%e17%). One NH only utilized Appendix 1 further elaborates on intervention quality aspects, and
CRP POCT during office hours because they saw too little potential also on focus group findings regarding CRP POCT utilization. Main
for CRP POCT to affect management decisions during their out-of- focus group findings regarding CRP POCT implementation knowledge
hours shifts. This NH remained content with this choice during are provided below.
the trial.
In Figure 1, the combined total use of CRP POCT is shown for all
participating NHs, which shows use of CRP POCT during the run-in Implementation Knowledge
period and continued use during the trial. In general, a pattern of
slightly increased use in winter months and a peak in July 2019 can be Figure 2 depicts the conceptual framework we developed based on
seen. A trough in CRP POCT use from April until July or August was identified subthemes and factors. Most factors can be considered both
more pronounced in some locations (data not shown). For 3 of total 18 facilitator and barrier (ie, if not well established, a factor can be
locations measurements prior to the last 100 measurements were considered a barrier). Central barriers/facilitators were getting people
missing due to device memory capacity. The imputed data based on on board, creating initial momentum, continued attention/enthu-
supply orders for this period show that these 3 locations performed a siasm, routine practice, and trust. A willingness to change practice and
large portion of measurements in the total of measurements. The broad recognition of importance helped getting NH staff on board.
average use per device during the trial was on average 3.1 times per This broad support was perceived necessary for the required changes
month (range 0.2e9.6). We do not have information on the proportion and time investment. Besides getting the people on board who will be
of tests that resulted in error messages, which either required to retest directly involved with CRP POCT, support on the management level
the sample or to collect a new sample. Among questionnaire re- was also mentioned. Management was perceived to facilitate imple-
spondents, POCT-users indicated that their workflow in case of an mentation if they were at least neutral toward this innovation and if
error message was to use a new sample (n ¼ 3), to retest the sample or the involved staff were allowed room for innovation.

Fig. 1. Combined total use of CRP POCT among all participating intervention group NHs. For 3 locations of in total 2 NHs, data is missing before Q3 2019 (n ¼ 2) and Q1 2019 (n ¼ 1).
An estimation for this missing data is made by averaging the order quantity between order dates and adding it to the total number of measurements.
972 T.M. Boere et al. / JAMDA 23 (2022) 968e975

Fig. 2. Conceptual framework of barriers and facilitators for CRP POCT implementation in the NH setting.

The nature and demandingness of innovation was related to get- that is what happened less in location 2.” [.] Nurse 2: “I think it maybe
ting people on board as well as to creating initial momentum. The also relates to, nurse 1 and 3, that physicians are also in your office, and
nature of innovation was considered comparable to other NH in- then you can just talk about it, and then it is, the lines of communication
novations, which facilitated implementation. Also, few barriers to are shorter, compared to when you have to track each other down.”
implementation were experienced because of perceived low demand Lastly, a process factor positioning of CRP POCT within current
and temporality of innovation (ie, for the duration of the trial). practices was brought into context of routine practice and trust.
Conversely, the complexity of test procedures was perceived as an Focus group participants seemed to experience fewer barriers to
initial threshold. However, this threshold was considered inherent to measuring CRP via POCT compared with send-out laboratory mea-
innovation and important to get past for achieving skilled and surement, in terms of shorter time-to-results, improved reach of
consequently routine practice. Enabling factors were a positive first smaller locations, night-/weekend-shift coverage, and in-house
experience, a short time between training and first experience, and management.
the instruction manual. Physician 2: “Yes, well, look, at ehmm. the lab visits smaller locations 1
Physician 1: “Well, for all new things you have to just get past some to 2 times a week. And then you have to pretty well move heaven and
sort of barrier, I shall say. We were properly instructed how to go about earth to ehh to, to get a CRP. And the results are also sometimes difficult to
it, but it took a few weeks before I saw my first patient. However, all, I, track down; I have to say, in practice. So the lines of communication are
the instruction manual was very clear. And then it was actually fun very short if you do have point-of-care.”
to do.” Conversely, for NHs that customarily use send-out laboratory rush
Process facilitators in the relation between continued attention/ orders, CRP POCT did not add value to current practice except for
enthusiasm and routine practice were proximity, 24/7 coverage, in- during nights or weekends. Furthermore, some NH employees
ternal task division, and role of external partners. Close proximity expressed less trust toward CRP POCT because of experienced de-
(between personnel as well as to the POCT-analyzer) and well-defined viations of POCT from laboratory measurements. These deviations
roles facilitated the process in case different personnel performed the were cautiously related to the extent of skilled measurement, which
measurement respectively were involved with patient selection and underlines the importance of adequate training for routine practice
follow-up. A perception regarding internal task division was that the and trust.
nurse should be able to suggest doing POCT but the decision should lie Factors identified within the subtheme implementation contents
with the physician. Furthermore, organizational and technical support involved experiences with instructions, practice, instruction tools,
from the laboratory, and nudges and incentives from the research recollection, POCT-user group, and training of new employees, which
team appeared to facilitate continued attention/enthusiasm and could be related to the process factors from initial momentum to trust.
routine practice. Physicians considered all training instructions satisfyingly clear, but
Nurse 1: “We did suggest the physicians, do you want to test for CRP? expressed the wish for conciseness in general. Practicing on each other
Ehm, and the physicians said do you want to test for CRP. You know, if you was perceived as a good training method and necessary investment to
actively draw each other’s attention to it then it happens more often, and obtain skilled practice. Practicing required many materials but also
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 973

revealed technical difficulties, which made POCT-users aware of their Intervention Quality
initial underestimation of required skills. The instruction manual
emerged as an important tool as recollection decreased with a longer The quality assessment indicates sufficient internal and external
time before seeing the first patient. validity of the intervention. The medical training attendance was on
Physician 3: “My name is XX, elderly care physician, and I have some average low, but training alternatives and public guideline availability
experience with it [CRP POCT]. A couple of times I could delegate it. When offered compensation to attain required knowledge. Physicians did
I had to do it myself one time, I had to rediscover for a moment how to go prefer a live session, which also served to build trust toward adjusting
about it, because there are actually many steps in the test procedure, and management decisions. In future implementation processes without
materials that are in the fridge and those that are not in the fridge, so it external oversight, we recommend to appoint a team champion
was nice to have a good instruction manual. But you did need to take the among physicians responsible for the medical training knowledge
time and ease to do it [for the first time].” within the group. A team champion would also contribute to oversight
Success factors for POCT-user group selection were group diversity, of POCT-user group composition and training status. The technical
manageable group size and holding on to this. The largest role was training of new employees appeared well-organized with in-house
ascribed to practice nurses during office hours and/or general nurses POCT trainers. Also, questionnaire respondents felt capable in CRP
for achieving 24/7 coverage. A back-up role for physicians seemed POCT utilization. Yet, the increased group size during the trial and
sufficient in both cases. For instance, during nights or weekends, the consequential infrequent use per POCT-user for some locations indi-
physician sometimes performed POCT to relieve the workload of the cate superfluity in POCT-user group size, which could have affected
out-of-hours chief nurse. Less oversight of the training of new em- routine practice.
ployees was related to somewhat more unrestricted or unwarranted Other parameters indicated adequate fidelity. Patient selection for
use of CRP POCT. CRP POCT generally corresponded with guideline recommendations
The training of new employees seemed organized well if the and CRP POCT use continued over time. In general, lower CRP POCT
training was part of the general new employees onboarding process, use in smaller-sized NH locations (ie, <100 residents) coincided with a
or if it was a prerequisite for the use of CRP POCT. The training of the similarly low number of participants, therefore, we assume low LRTI
in-house POCT-trainer was perceived as satisfactory and clear, and the incidence rather than insufficient CRP POCT implementation.
method of coaching was deemed suitable. Instances of less organized
training methods were characterized by high staff turnover and Barriers and Facilitators
coincided with a generally lower recollection of training and
decreased attention/enthusiasm. Consequently, new physicians dele- Because the trial results support the benefit of large-scale imple-
gated or did not use POCT-measurement. Reportedly, the training of mentation of CRP POCT, we provide recommendations (Figure 3) for
new employees could improve if physicians stimulate and frequently future implementation processes based on barriers/facilitators iden-
interact with the in-house POCT-trainer. tified in the current process evaluation.31
A willingness to change practice and perceived added value of the
Discussion innovation among staff similarly appeared as facilitators in other NH
process evaluation studies with different types of in-
The aim of this mixed-methods process evaluation was to assess terventions.21,23,32,33 The perceived low-demanding nature of CRP
the implementation and use of CRP POCT in NHs. We evaluated POCT implementation seemed to improve initial commitment. Other
intervention quality, and assessed barriers and facilitators for the NH process studies also describe this relation between demanding-
implementation and use of CRP POCT. This process evaluation showed ness of innovation and its uptake, which suggests a general point for
sufficient intervention quality and together with implementation consideration in implementation processes.23,32e34 Likely, the
knowledge these findings provide clues and conditions for logistics of demandingness of the innovation should be proportional to its
CRP POCT implementation in NHs. Trial results showed that access to perceived added value.
CRP POCT reduced antibiotic prescribing for suspected LRTI and, Facilitators for continued attention/enthusiasm and skilled routine
therefore, support large-scale implementation.31 practice were 24/7 coverage with a diverse and coordinated POCT-user

Fig. 3. Key recommendations for CRP POCT implementation in the NH setting.

974 T.M. Boere et al. / JAMDA 23 (2022) 968e975

group, and proximity, (ie, short lines of communication between analysis, interpretation, or in writing the manuscript or authority over
personnel and short distance to the POCT-analyzer). This interaction any of these activities.
between continued attention, proximity, and routine practice was also
seen in another study.33 POCT-users experienced the instruction
References
manual as useful, especially in case of waned practice. Furthermore,
other studies also describe staff turnover as a barrier to group medical 1. Dhawan N, Pandya N, Khalili M, et al. Predictors of mortality for nursing home-
training knowledge.21,23,34 Although communication logs indicated acquired pneumonia: A systematic review. Biomed Res Int 2015;2015:285983.
difficulty for some NHs to retain group knowledge in the face of high 2. Lim CJ, Kong DC, Stuart RL. Reducing inappropriate antibiotic prescribing in the
residential care setting: Current perspectives. Clin Interv Aging 2014;9:
staff turnover, this was mentioned to a lesser extent in the focus group 165e177.
interviews. We recommend regular discussion of experiences with 3. Lim CJ, McLellan SC, Cheng AC, et al. Surveillance of infection burden in resi-
CRP POCT among staff as part of group knowledge building and for dential aged care facilities. Med J Aust 2012;196:327e331.
4. Loeb M, Simor AE, Landry L, et al. Antibiotic use in Ontario facilities that pro-
safeguarding continued attention.
vide chronic care. J Gen Intern Med 2001;16:376e383.
Physicians agreed that patient selection for CRP POCT required the 5. Mills K, Graham AC, Winslow BT, Springer KL. Treatment of nursing home-
potential for therapeutic consequences. This finding concurs with acquired pneumonia. Am Fam Phys 2009;79:976e982.
guideline recommendations and other studies involving CRP 6. Hollaar V, van der Maarel-Wierink C, van der Putten GJ, et al. Defining char-
acteristics and risk indicators for diagnosing nursing home-acquired pneu-
POCT.33,35e37 Another perceived value of CRP POCT, however, was for monia and aspiration pneumonia in nursing home residents, using the
prognostication and monitoring. As our trial was not designed for electronically-modified Delphi Method. BMC Geriatr 2016;16:60.
assessment of these purposes, possibly, future studies could focus on 7. Loeb M, Carusone SC, Goeree R, et al. Effect of a clinical pathway to reduce
hospitalizations in nursing home residents with pneumonia: A randomized
CRP-values related to prognosis and monitoring along the disease controlled trial. JAMA 2006;295:2503e2510.
trajectory.11,36,37 Some studies described CRP POCT as a “learning tool 8. Porfyridis I, Georgiadis G, Vogazianos P, et al. C-reactive protein, procalcitonin,
to detect patients needing antibiotics,” through improved clinical clinical pulmonary infection score, and pneumonia severity scores in nursing
home acquired pneumonia. Respir Care 2014;59:574e581.
pattern recognition.37,38 Similar to other studies, physicians in this 9. van Buul LW, van der Steen JT, Doncker SM, et al. Factors influencing antibiotic
study believed that CRP POCT reinforced their diagnostic toolkit, while prescribing in long-term care facilities: A qualitative in-depth study. BMC
they were also cautious toward over-reliance.36,38 Geriatr 2014;14:136.
10. Andreeva E, Melbye H. Usefulness of C-reactive protein testing in acute cough/
respiratory tract infection: An open cluster-randomized clinical trial with C-
Strengths and Limitations reactive protein testing in the intervention group. BMC Fam Pract 2014;15:80.
11. Arinzon Z, Peisakh A, Schrire S, Berner Y. C-reactive protein (CRP): An impor-
tant diagnostic and prognostic tool in nursing-home-associated pneumonia.
A strength of this process evaluation was the use of a mixed- Arch Gerontol Geriatr 2011;53:364e369.
methods design. Although the use of multiple data sources strength- 12. Cals JW, Butler CC, Hopstaken RM, et al. Effect of point of care testing for C
reactive protein and training in communication skills on antibiotic use in lower
ened our findings, the questionnaire response rate was limited (33%). respiratory tract infections: Cluster randomised trial. BMJ 2009;338:b1374.
The use of focus group interviews enriched implementation knowl- 13. Minnaard MC, de Groot JAH, Hopstaken RM, et al. The added value of C-reactive
edge regarding CRP POCT. Of note, because we evaluated in study protein measurement in diagnosing pneumonia in primary care: A meta-
analysis of individual patient data. CMAJ 2017;189:E56eE63.
context, certain factors need to be considered for implementation
14. Nouvenne A, Ticinesi A, Folesani G, et al. The association of serum procalcitonin
outside the research setting. As aforementioned, in the absence of and high-sensitivity C-reactive protein with pneumonia in elderly multimorbid
external oversight from a research team, NHs could appoint a team patients with respiratory symptoms: Retrospective cohort study. BMC Geriatr
champion and partner with a laboratory. Furthermore, this process 2016;16:16.
15. van Vugt SF, Broekhuizen BD, Lammens C, et al. Use of serum C reactive protein
evaluation precedes the current severe acute respiratory syndrome and procalcitonin concentrations in addition to symptoms and signs to predict
coronavirus 2 pandemic. Consequently, our findings do not consider pneumonia in patients presenting to primary care with acute cough: Diag-
the additional strain on health care personnel and doubtlessly nostic study. BMJ 2013;346:f2450.
16. Hopstaken RM, Kleinveld HA, van Balen JAM, et al. Guideline for point-of-care
increased difficulty for implementation processes. testing (POCT) in general practice (guideline in Dutch). The Dutch College of
General Practitioners (NHG) 2015:1e18. Available at: https://
www.nhg.org/sites/default/files/content/nhg_org/uploads/final_richtlijnpoct_
Conclusions 2015lmlres.pdf. Accessed October 19, 2021.
17. Butler CC, Gillespie D, White P, et al. C-Reactive Protein Testing to Guide
This process evaluation showed sufficient quality of providing CRP Antibiotic Prescribing for COPD Exacerbations. N Engl J Med 2019;381:
111e120.
POCT in Dutch NHs. We processed findings of intervention quality and 18. Cals JW, Ament AJ, Hood K, et al. C-reactive protein point of care testing and
implementation knowledge into key recommendations for CRP POCT physician communication skills training for lower respiratory tract infections
implementation in this setting, such as short distances and short lines in general practice: Economic evaluation of a cluster randomized trial. J Eval
Clin Pract 2011;17:1059e1069.
of communication. Future research could focus on CRP POCT use (eg, 19. Little P, Stuart B, Francis N, et al. Effects of internet-based training on antibiotic
logistics and role in management decisions) in countries with prescribing rates for acute respiratory-tract infections: A multinational, cluster,
different organization of care in NHs, for instance, without in-house randomised, factorial, controlled trial. Lancet 2013;382:1175e1182.
20. Boere TM, van Buul LW, Hopstaken RM, et al. Using point-of-care C-reactive
physicians or with different practices regarding hospital referrals. protein to guide antibiotic prescribing for lower respiratory tract infections in
elderly nursing home residents (UPCARE): Study design of a cluster random-
ized controlled trial. BMC Health Serv Res 2020;20:149.
Acknowledgments 21. Appelhof B, Bakker C, van Duinen-van den IJssel JCL, et al. Process evaluation of
an intervention for the management of neuropsychiatric symptoms in young-
We thank the physicians and nurses from nursing homes partici- onset dementia. J Am Med Dir Assoc 2018;19:663e671.
22. Mentes JC. Tripp-Reimer T. Barriers and facilitators in nursing home inter-
pating in the UPCARE trial who responded to our questionnaire and vention research. West J Nurs Res 2002;24:918e936.
who participated in our focus group interviews. We thank the research 23. Zwijsen SA, Smalbrugge M, Eefsting JA, et al. Grip on challenging behavior:
team member Melissa van Essen for her contribution to the data Process evaluation of the implementation of a care program. Trials 2014;15:
302.
collection and first part of open coding of focus group interviews. We
24. Leontjevas R, Gerritsen DL, Koopmans RT, et al. Process evaluation to explore
thank Jochen Cals and Ruth Veenhuizen for their contribution to the internal and external validity of the "Act in Case of Depression" care program
conception and initial design of the UPCARE study. We also thank in nursing homes. J Am Med Dir Assoc 2012;13:488.e1e488.e8.
Saltro diagnostic center location Utrecht for their collaboration and 25. Linnan L, Steckler A. Process evaluation for public health interventions and
research: An overview. In: Linnan L, Steckler A, editors. Process Evaluation for
quality control oversight of CRP POCT-use during the trial. The funding Public Health Interventions and Research. San Francisco, CA: Jossey-Bass; 2002.
bodies had no role in the design of the study, data collection, data p. 1e23.
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 975

26. Dutch Association of Elderly Care Physicians (Verenso). Lage luchtweginfecties 32. Anrys P, Strauven G, Roussel S, et al. Process evaluation of a complex inter-
bij kwetsbare ouderen (Guideline Respiratory Tract Infections). Utrecht, the vention to optimize quality of prescribing in nursing homes (COME-ON study).
Netherlands: Verenso. 2018. Available at: https://www.verenso.nl/richtlijnen- Implement Sci 2019;14:104.
en-praktijkvoering/richtlijnendatabase/lage_luchtweginfecties. Accessed 33. Eley CV, Sharma A, Lecky DM, et al. Qualitative study to explore the views of
October 19, 2021. general practice staff on the use of point-of-care C reactive protein testing for
27. Minnaard MC, van de Pol AC, Broekhuizen BD, et al. Analytical performance, the management of lower respiratory tract infections in routine general
agreement and user-friendliness of five C-reactive protein point-of-care tests. practice in England. BMJ Open 2018;8:e023925.
Scand J Clin Lab Invest 2013;73:627e634. 34. Pieper MJC, Achterberg WP, van der Steen JT, Francke AL. Implementation of a
28. Minnaard MC, van de Pol AC, de Groot JA, et al. The added diagnostic value of Stepwise, Multidisciplinary Intervention for Pain and Challenging Behaviour in
five different C-reactive protein point-of-care test devices in detecting pneu- Dementia (STA OP!): A process evaluation. Int J Integr Care 2018;18:15.
monia in primary care: A nested case-control study. Scand J Clin Lab Invest 35. Hardy V, Thompson M, Keppel GA, et al. Qualitative study of primary care
2015;75:291e295. clinicians’ views on point-of-care testing for C-reactive protein for acute
29. Matheeussen V, Van Hoof V, Loens K, et al. Analytical performance of a plat- respiratory tract infections in family medicine. BMJ Open 2017;7:e012503.
form for point-of-care CRP testing in adults consulting for lower respiratory 36. Kuil SD, Schneeberger C, van Leth F, et al. "A false sense of confidence" The
tract infection in primary care. Eur J Clin Microbiol Infect Dis 2018;37: perceived role of inflammatory point-of-care testing in managing urinary tract
1319e1323. infections in Dutch nursing homes: A qualitative study. BMC Geriatr 2020;20:
30. van der Steen JT, Ooms ME, Ader HJ, et al. Withholding antibiotic treatment in 450.
pneumonia patients with dementia: A quantitative observational study. Arch 37. Phillips R, Stanton H, Singh-Mehta A, et al. C-reactive protein-guided antibiotic
Intern Med 2002;162:1753e1760. prescribing for COPD exacerbations: A qualitative evaluation. Br J Gen Pract
31. Boere TM, van Buul LW, Hopstaken RM, et al. Effect of C reactive protein point- 2020;70:e505ee513.
of-care testing on antibiotic prescribing for lower respiratory tract infections in 38. Cals JW, Chappin FH, Hopstaken RM, et al. C-reactive protein point-of-care
nursing home residents: Cluster randomized controlled trial. BMJ 2021;374: testing for lower respiratory tract infections: A qualitative evaluation of ex-
n2198. periences by GPs. Fam Pract 2010;27:212e218.
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 975.e1

Supplementary Data

Supplementary Table 1
Experienced Relevance, Satisfaction, and Feasibility of CRP POCT in the NH Setting

Statements Response (n) Mean Score, Range*

I am satisfied with the method of the medical training session 23 7.1, 4e9
I am satisfied with the contents of the medical training session 23 7.2, 4e9
I am satisfied with the method of the technical POCT training session 21 7.5, 6e8
I am satisfied with the contents of the technical POCT training session 11 7.8, 7e9
I feel capable/adept in the use of CRP POCT 21 7.2, 3e10
I feel capable/adept in the interpretation of CRP POCT results 29 7.6, 4e10
I use the CRP cut-off values as proposed by the NH LRTI guideline for my management decisions 19 8.1, 5e10
I find the CRP cut-off values as proposed by the NH LRTI guideline useful in clinical practice 19 8.0, 5e10
I trust CRP POCT results in my management decisions 19 8.2, 6e10
I feel that CRP POCT has added value to my management decisions 19 8.5, 7e10
I would have acted differently in my management decisions had I not used CRP POCT 19 7.2, 3e10
I would recommend the use of CRP POCT to colleagues 19 8.6, 7e10
I think that the POCT-analyzer is user-friendly 23 7.8, 5e10

*Respondents answered to each statement with an agreement score between 1 (totally disagree) e 10 (totally agree).

Supplementary Table 2
Experienced Influences on Antibiotic Prescribing

Statement: “the following Response (n) Mean

factors have changed my antibiotic Score,
prescribing for LRTI the past year” Range*

The use of CRP POCT 19 7.6, 6e9

The (new) LRTI guideline for the NH setting 19 7.3, 5e9
The medical training sessiony 15 4.6, 1e8
Participation in the UPCARE study in general 19 6.7, 2e9
Other influences 19 5.2, 1e9

*Respondents answered to the statement for each factor with an agreement

score between 1 (totally disagree) e 10 (totally agree).
y
Answers for the influence of the medical training session were omitted if the
respondent indicated that he/she did not follow any kind of medical training for the
UPCARE study.
975.e2 T.M. Boere et al. / JAMDA 23 (2022) 968e975

Supplementary Table 3
Feasibility and Accessibility of CRP POCT in the NH Setting

Statements Answer Response (n) n (%)

I feel that I have the time to use CRP POCT at initial consultation Yes 25 22 (88)
No 3 (12)
I think that CRP POCT can be used well during nights and weekends. Yes 25 20 (80)
No 5 (20)
I think that the location of the POCT-analyzer is properly accessible for its users Yes 27 23 (85)
No 1 (4)
N/A 3 (11)
I know where to find CRP POCT materials Yes 27 22 (82)
No 3 (11)
N/A 2 (7)
I know where to find CRP POCT instruction materials Yes 27 23 (85)
No 4 (15)

Appendix 1 Patient selection for CRP POCT

Views concurred with guideline recommendations with regard to
This appendix provides supplementary results regarding inter- patient selection, for instance, physicians thought that CRP POCT
vention quality and implementation knowledge. should only be used if there are potential therapeutic consequences.
Differing views appeared for patient selection among psychogeriatric
patients. Some physicians experienced no differences in patient se-
Intervention Quality
lection for POCT, while others experienced differences in infection
recognition among psychogeriatric compared to somatic patients. The
Satisfaction, relevance, and feasibility
patient’s understanding was viewed as a barrier for performing POCT,
Satisfaction scores were on average between 7 and 8 (on a 10-point
but the finger prick was perceived not to be more burdensome. Broad
scale) for methods and content of the POCT and medical training
laboratory assessment was preferred over CRP POCT for severely ill
sessions (see Supplementary Table 1). Respondents who received the
patients receiving empirical antibiotic treatment.
medical training from a colleague or via instruction materials in
absence of a live session were less satisfied with the method (score:
Measurement
6.6) and content (score: 6.4). Respondents felt to a certain extent
The POCT-analyzer was experienced as clear, user-friendly, quick,
capable in CRP POCT performance (score: 7.2) and interpretation
and pleasant. The instructions were thought to be straight forward;
(score: 7.6). Physicians often communicated that they perceive CRP
however, the number of test procedure steps and required technical
POCT as a simple or normal procedure, analogous to glucose mea-
skills necessitated sufficient practice and the use of the instruction
surement and other measurements requiring a finger prick.
manual. The finger prick measurement was thought to be similar to
Respondents experienced added value of CRP POCT and trust in
other tests such as blood glucose measurement. Lastly, a desired
CRP POCT results for management decisions (scores: 8.5 respectively
improvement of a POCT such as for CRP was the ability to differentiate
8.2). Users scored 7.2 (range 3e10) when asked if CRP POCT had
between viral and bacterial causes.
changed their management decisions. Supplementary Table 2 shows
that respondents thought that the use of CRP POCT and the NH LRTI
Evaluation
guideline influenced their antibiotic prescribing most (scores: 7.6
Short lines of communication facilitated retrieval of CRP POCT re-
respectively 7.3) and the medical training session the least (score: 4.6).
sults for evaluation. Doubts were raised about telephone communi-
Other influences on prescribing decisions that were mentioned among
cation of results to out-of-hours service physicians without them
questionnaire respondents were increased general attention for the
seeing the patient. In that case, CRP POCT might unjustifiably have
issue of antimicrobial resistance (n ¼ 4), medical specialist education
more weight in treatment decisions.
or peer-review (n ¼ 3), and the regular staff meetings regarding
The guideline-specified cut-off values for CRP in the treatment
pharmacotherapy (n ¼ 2).
decision were perceived as adequate. A side note was made that the
guideline should leave room for individual variation. CRP-values were
Feasibility in instances unexpected, either surprisingly low or high. Most often
the CRP-value was surprisingly low, which justified the choice for no
Supplementary Table 3 shows that most (22 of 25) respondents felt antibiotics and eased the choice for watchful waiting.
like they had the time to use CRP POCT around initial consultation. The role of CRP POCT in treatment decisions was described in
Also, 20 of 25 respondents thought CRP POCT to be appropriate for varying degrees. Instances of larger dependency were described as
nights or weekends. The POCT-analyzer, materials, and instructions support for not prescribing antibiotics, a confirmation of gut feeling, or
were thought to be easily accessible (>80% of 27 respondents) and differentiation of infection in case of unclear disease focus. Also, CRP
user-friendly (score: 7.8). Respondents who included participants POCT was found useful for clear and quick guidance on (short-term)
indicated actual use and usefulness of LRTI guideline-proposed cut-off management, specifically in initial phases of disease. In case of lower
values in clinical practice (scores: 8.0 respectively 8.1). dependency, a greater importance was ascribed to the clinical signs
and symptoms, along which CRP merely offers a small and indirect
Implementation Knowledge course of direction.
Physician 4: “And also, in my opinion, you of course have to give
Utilization of CRP POCT priority to your clinical impression [.] you are treating a patient, not a
General experiences with the use of CRP POCT were shared number.”
regarding patient selection for CRP POCT, measurement and evaluation Experienced effects of CRP POCT were a direct impact on antibiotic
of results. prescribing (ie, fewer), having a more justified management decision,
 T.M. Boere et al. / JAMDA 23 (2022) 968e975 975.e3

an ease of mind, but also instances of limited effects on antibiotic Physician 5: “You do have something ‘extra’ [with CRP POCT], also
prescribing. Trust in CRP POCT increased if spontaneous recovery was sometimes it is reassuring to see that the CRP-level is low. Then you do
observed after withholding antibiotics. sleep better, I shall say.”