Due to nursing shortages, an ageing population and increasing care demand, there is a growing interest in parenteral medication administration at home (PMAaH), comprising the administration of parenteral medication in the home situation of patients. The operational design of such PMAaH care pathways is complex, resulting in many variations of adoptions, showing a need for a quality framework. Although quality indicators (QIs) have been proposed to monitor the quality of specific care pathways, a generic quality framework for all types of PMAaH is lacking. Therefore, this study proposes a generic quality set for PMAaH, which includes structure and process QIs, to benchmark and redesign PMAaH care pathways to ensure high quality.

A generic QI set was developed for PMAaH using a systematic RAND appropriateness method adapted at the third phase. This method consisted of a scoping review to identify indicators, an expert panel rating phase including an online questionnaire and subsequent panel meeting to assess the appropriateness of the indicators and a retrospective practice testing to evaluate the feasibility, clarity and measurability of the indicators. After the practice testing, which consisted of an online questionnaire where experts could indicate the implementation state of all indicators in their hospital, a third expert panel adjusted the set to increase the likelihood of implementation in practice.

The experts, all healthcare professionals involved in PMAaH processes, were recruited using the snowball sampling technique from three large Dutch, teaching hospitals. Subsequently, a practice testing by self-assessment was conducted in seven large Dutch teaching hospitals.

17 and seven healthcare professionals with diverse backgrounds participated in the online questionnaire and panel meeting, respectively.

This study proposes a generic quality set for PMAaH that hospitals can use to redesign and benchmark PMAaH care pathways to assure high quality. The practice testing confirmed that there is a need for this structured quality set.

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

In recent years, healthcare providers have increasingly shown interest in value-based care initiatives that move care from in-hospital settings to care settings closer to or in patients’ homes, such as hospital at home1 and outpatient parenteral antimicrobial therapy (OPAT).2 This shift is further fuelled by, among others, the COVID-19 pandemic, experienced nursing shortages and an ageing population with increasing healthcare demands, all of which create a need to use hospital capacity more efficiently.

One of the opportunities to reduce in-hospital care is parenteral medication administration at home (PMAaH). PMAaH is the home-based administration of intravenous, intramuscular and subcutaneous medication. Besides creating hospital capacity, this has also proven to increase patient satisfaction3 4 and to be cost-effective compared with inpatient medication administration4 5 for a range of patient types. In the Netherlands, many hospitals have adopted PMAaH pathways, with the most common being infectious patients requiring infusion treatment with antibiotics and oncology patients requiring immunotherapy. During this treatment at home, the medical specialist of the hospital is responsible for the patients, but the administration at home can be executed by hospital-employed nurses or third parties, such as home healthcare organisations.

The organisation of PMAaH is inherently complex and multidisciplinary. Multiple healthcare professionals, such as physicians, nurses, pharmacists and pharmacy technicians, are involved in various stages of the care pathway, including enrolment of patients in the programme, medication preparation, administration, follow-up and discontinuation. Since PMAaH pathways were historically disease-specific or drug-specific (eg, OPAT), many variations in PMAaH pathway applications have emerged, even within a single hospital or region. Hence, a multimorbid patient may receive parenteral medication administered by a hospital nurse at home, who brings the medication from the hospital pharmacy, as well as by a home-healthcare nurse, after medication is delivered by commercial parties at the patient’s home. These variations in PMAaH pathways may lead to unwanted variations in the quality of the provided care.

Given the complexity of PMAaH pathways and the variations in the applications of PMAaH, the measurement and benchmarking of a generic set of quality indicators (QIs) can provide insights into the quality and safety of the adoption of care processes, as seen in mental health6 and palliative care.7 These QIs are used to evaluate healthcare quality in terms of the characteristics and the context in which care is provided (ie, structure QIs), and processes and actions that are taken in the care pathway (ie, process QIs).8 Both structure and process QIs provide insight into the differences in medication administration practices, allow for monitoring and evaluation of the adherence to standardised protocols, guidelines and best practices and enhance collaboration between healthcare providers.9 Therefore, these QIs intend to measure to what extent care processes are implemented in line with the current professional standards to guarantee the quality of care processes9–11 and guide improvements in PMAaH pathways that in turn improve patient health outcomes.

Several QI sets are proposed and evaluated for subsets of PMAaH. For example, Berrevoets et al2 and March-López et al12 developed and tested a set of QIs for assessing the appropriateness of OPAT care at home. Furthermore, Kirwin et al,13 and King et al14 developed process indicators for the activities of pharmacists. However, these sets are all developed for a specific care pathway related to a particular patient group or medication type (eg, antibiotics), or related to a specific profession, whereas PMAaH applications typically encompass the variety of these applications. Moreover, currently no generic set of QIs for PMAaH is available. In this study, we therefore aim to develop and pilot test a generic set of structure and process QIs to be used by hospitals in the (re)design and quality evaluation of PMAaH pathways. The aim of this set is to complement the already existing quality frameworks for parenteral medication in the hospital setting, since this set offers a framework for the outpatient setting, not including safety concerns related to specific medications, but is generic for all parenteral medication at home.

We used a systematic, RAND appropriateness method (RAM) to develop a set of generic QIs for PMAaH pathways,15 16 which we slightly adapt to the context of process and structure QIs. This RAM method combines the commonly used Delphi method, which focuses on independent judgements of experts using questionnaires, with focus groups, where an expert panel discusses their judgements.17 18 We chose this method to ensure that the QIs should be appropriate for the PMAaH setting, since it promotes finding consensus, which is especially important in a multidisciplinary setting such as the PMAaH pathways.

Figure 1 visualises the methodology in three phases, adapted from Fitch et al.15 During the first phase, we collected relevant articles on quality assurance of PMAaH pathways by a scoping review. In the second phase, we reached consensus among an expert panel, starting with an online questionnaire, without interaction and followed by panel meeting discussions. If no consensus was reached in the questionnaire round, the QIs were discussed in a panel meeting with various healthcare professionals involved in PMAaH. During the last phase, multidisciplinary teams retrospectively assessed to what extent the acquired QIs were already implemented in practice, by scoring their hospital’s policies on the indicators in a practice testing. The methodology of this study follows the RAM as proposed by Fitch et al,15 except for the last phase. Fitch et al15 propose a retrospective or prospective evaluation of the outcomes using clinical records or clinical decision aids. Since, in this study, process and structure QIs were developed, we aim to retrospectively assess the extent of implementation of these QIs in the last phase, rather than the appropriateness using clinical records.

The study ran between February 2023 and March 2024. The second phase (expert panel ratings) of the method was carried out in three large Dutch teaching hospitals and the last phase (practice testing) in seven Dutch teaching hospitals. The three initial hospitals were selected because they all have multiple PMAaH pathways applied in practice and therefore benefit from a generic QI set to redesign their pathways and evaluate their quality. Additionally, these hospitals are all members of the mProve network, creating the opportunity to test this set in seven similar hospitals during the third phase. Ethics approval was provided on 24 April 2023 by the University of Twente-BMS Domain Humanities and Social Sciences ethics board, registration number 230 125. Reporting of this study was according to the ACcurate COnsensus Reporting Document guideline.19 Since this study focuses on structure and process indicators, and not outcome indicators, there was no patient or public involvement in this study. The study was not registered prospectively.

This section presents the QIs identified by the scoping review, the modification of the QIs by the expert panel rating phase and the final QI set of PMAaH after the retrospective practice testing.

We developed a generic set with 32 QIs for PMAaH using a slightly modified RAM. The practice testing showed that it is possible to monitor the quality of both procedural and structural aspects of PMAaH pathways and that there is a need for such a set, since many QIs agreed on by directly involved experts are not yet implemented in large teaching hospitals in the Netherlands. Our research shows that many QIs are only implemented on a care pathway level, leading to variations. This set could be used both for benchmarking and quality improvement. To the authors’ knowledge, this study is the first to present a coding scheme for the data analysis of a panel meeting within the RAM setting. Although the outcomes of a panel meeting within the RAM are often clear, this coding scheme adds the motivation behind these decisions, which deepens the discussion of the results.

Our study has strengths and weaknesses throughout all phases of the RAM method. In the scoping review, a highly relevant article from Berrevoets et al2 was only discovered via snowballing, which suggests that other studies might have been missed due to the used search strings. However, since only one of the QIs (final QI P12 regarding availability of rescue medication) was extracted from a single study, and all other QIs are based on multiple studies, we expect that no relevant QIs are missed.

In the expert panel ratings phase, both the Dutch and English translations were shown, but since Dutch was the primary language of the questionnaire and panel meeting, this could have an impact on the assessment of the experts. We have tried to minimise this impact through a grammar and spelling check by a native speaker for both sets. During the first expert panel ratings round, several attendees warned that they perceived the questionnaire as too lengthy and time consuming. This could have led to response fatigue, resulting in respondents being less motivated to assess the last QIs as effective and critical as the first QIs.25 Our results do suggest that this might have played a role, since after the process QI LP8 (regarding the patient and family information details), no QI was labelled uncertain or inappropriate, while during practice testing, many of these indicators required adaptation. To overcome this in the future, we suggest only presenting the QIs as a whole with a remark option for the included sub-QIs, to reduce the length of the questionnaire. We do not believe this response fatigue has led to biased results in the end, since the practice testing ensured that all experts considered all QIs once more (besides the panel meeting) and were perceived to be more critical during practice testing than in the first round of stakeholder consensus. This showcases the importance of practice testing. The response fatigue did not lead to internal invalidity of the entire QIs compared with their sub-QIs, since there were no QIs that were labelled appropriate while all sub-QIs were labelled inappropriate or vice versa.

During the second expert panel rating round, time was a restricting resource, since it required multiple healthcare professionals from multiple hospitals to be available at the same time. Because of the multicentre approach, we chose a virtual panel meeting, since it is reported that attendance is comparable to live meetings, see Halliday et al.26 In the panel meeting, there was sufficient time to discuss all uncertain labelled QIs, ensuring that all experts were able to express their opinion and come to consensus, but limited time to individually discuss all remaining QIs. Besides, the developed coding scheme added value to the extraction of the results in this second round, and we propose the validation of this scheme as a subsequent research step. We believe that the fact that the scheme was not validated yet did not impact our results, since only three (sub-)QIs were discussed, which all fitted in the coding scheme. It is worth mentioning that other panel meetings within the RAM could come up with arguments that do not completely fit in this coding scheme.

Most of the adaptations in the QIs resulted from the practice testing phase, which was not expected upfront. This indicates that it is hard to assess the feasibility of a QI until it is tested in practice. An example of this is structure QI S2, where it was stated that a PMAaH committee must include a physician experienced with each administered medication in the home situation, while it is not feasible to include a separate physician for each administered medication. Wollersheim et al27 shows that 10–20% of the indicators were not measurable, which was revealed in practice testing, which is in line with our study. While not all RAM studies include a practice testing phase at the end, our study shows that the set is updated when such a phase is included, which is in line with other literature that does include a practice testing phase.7 28 This result is important, not only because the finalised QI set is more feasible, clear and measurable and therefore more likely to be adopted in practice, but also because it shows that there is a need for further research on the impact of systematic methodologies that were deployed before the practice testing, in relation to the impact of practice testing itself. These findings are in line with the literature, as Malmqvist et al29 showed that a pilot study in practice has the potential to increase the quality of the research. On the other hand, these practice testing results question the validity of the results of the first two phases. We believe that these phases still hold validity, since the QIs that were entirely removed after the practice testing were removed because their aim was already included in standard medical practice or other QIs. We propose that further research is needed on why it is hard for expert panels to assess the feasibility, clarity and measurability of a QI in the second phase of the RAM.

The second and third phases of our research setup faced disengagement of invited experts, resulting in changing expert panel compositions. In the first round of the expert panel, all professions participated, whereas in the second round, due to unavailability and workload of the healthcare professionals, not all professions participated in the expert panel meeting, resulting in an under-representation of medical specialists in the second panel meeting, which could lead to a bias in the outcomes. In the third phase, hospital pharmacists were responsible for executing the practice testing, and due to the added workload, we did not require specific experts to join the multidisciplinary teams. The practice testing could therefore have a pharmaceutical bias towards the outcomes.

Because of the scope of our research design, not all aspects of PMAaH care pathways were considered. It is important to note that for specific medication types, specific QIs should be considered. For example, the safe handling of hazardous drugs should be considered for chemotherapy infusion treatments,30 while the organisation of laboratory results and stability testing should be considered for antibiotics infusion treatments.2 31 It is important to note that these aspects are already considered in treatment guidelines and only sometimes have to be slightly adjusted in an outpatient setting. Therefore, we believe that a generic set is feasible, combined with the existing guidelines. Note that in parenteral drug administration, incompatibilities and diluting are important concerns. In the context of this study in the Netherlands, there are nationwide standards available on this, to ensure standardisation across organisations. We recommend that hospitals consider at least alignment on these issues within their own care network.

Since the QI set is generic, it only includes structure and process indicators and no outcome indicators, since outcome indicators encompass the effects on the patient population and can be specific to the disease or drug type and are therefore of less relevance when developing generic PMAaH QIs. In this study, patients were not involved in the expert panel rating phase. This is often the case in the development of QI sets,32 and this specific set is primarily developed for quality improvement on a procedural and structural level. Therefore, the decision was made not to include patients, but only include experts that are responsible for providing the optimal preconditions to deliver good quality of care. When outcome indicators are included in the set as well, we believe that patient involvement is beneficial. Besides, when a treatment is relocated to the home situation of the patient, a hospital should also consider the quality of the aspects unrelated to parenteral medication, such as the definition of responsibilities when a patient leaves the hospital.

A close analysis of the inappropriately labelled QIs in the phases of the m-RAM provides insights into the development of a generic QI set. In the expert panel rating phase, two out of three QIs were labelled inappropriate, because they both required something that might not be feasible in all situations (ie, LP1: a clinician takes leadership of PMAaH team and LS16: a policy regarding the travel distance from patients to hospital). Since the expert panel discussed in the panel meeting that this would not affect the quality of care, they were removed from the set. For the other QI, the discussion in the panel meeting showed that there were two different interpretations of the QI. A panel meeting gives the opportunity to discuss both interpretations, after which the panel meeting decided that this was not necessary or already covered in another QI. It could be possible that for other QIs, there were also multiple interpretations possible, which did not have the possibility of discussion in a panel meeting. For future RAM research, we would propose to include a motivation for each QI based on the provided references, to prevent this.

The practice testing showed that the implementation of the QIs varies highly between hospitals, since there is only one QI that is fully implemented, and also two QIs that are not implemented at all. These latter QIs are related to self-administration, so it could be that this is not applicable for the questioned hospitals. For future research, it is therefore recommended to add a not applicable option in the practice testing, which was not included in this study. While recent studies show that self-administered OPAT is a feasible treatment option,33 34 the lack of self-administration options in hospitals could be related to negative financial incentives.35 During practice testing, experts often mentioned that some QIs were already arranged for by general protocols and/or treatment plans, although not specifically dedicated to administration in the home situation. Because of this, some QIs were removed after the practice testing (LP2 and LP19). Experts also remarked that many QIs were organised on a disease or medication type level, but not on a general level, which resulted in a partially met score on many QIs. Consequently, it is important to understand that a low percentage of full implementation does not directly imply unsafe care for patients.

The QIs that are implemented in the majority of the hospitals are on the operational level (eg, competence of caregiver that includes patients, policy on first dose administration and selection and removal of drug delivery device), while strategic level (both in process and structure) QIs (eg, structure programme for a framework of safe and effective care and system of ongoing quality assurance) are implemented in only a few hospitals. This practice testing shows that this study, which focuses on providing a generic framework for providing safe and effective PMAaH, is contributing to this lack of and need for structured programmes.

For the deployment of this QI set in practice, we have the following suggestions. Preferably, this set becomes the nation-wide quality standard, where a committee could assess hospitals using these QIs. Since this is not yet the case in the Netherlands, the deployment will consist of a self-assessment, in line with the practice testing. We propose that a self-assessment team consists of the professions discussed in QI 2: a physician, pharmacist and nurse, all experienced with PMAaH. The workload of this self-assessment, and afterwards improvement of the organisation, will highly vary per hospital, depending on which systems are already in place to monitor the PMAaH quality.

This QI set is not only useful for hospitals that want to evaluate and/or improve the quality of their PMAaH care pathways, but also for organisations that collaborate with hospitals, for example, the ones responsible for the administration at patients’ homes. It is common that hospitals collaborate with multiple home healthcare organisations and that home healthcare organisations are contracted by multiple hospitals. Home healthcare organisations report challenges with monitoring their performance.36 The generic QI set presented in this study could also ensure a standardisation in quality expectations for home healthcare organisations.

For the seven included hospitals, participating in this research resulted in practical implications on their organisation. All hospitals have expressed in the mProve working group that they are in the process of, or have already, set up a formally established committee dedicated to PMAaH policy, as stated in QI S2. Once such a committee is established, this committee is in charge of establishing the other QIs in practice. For future applications of this QI set, we envision this set to be a nationwide set for hospitals, where the assessment of the implementation of these QIs is not only informed through self-assessment.

For future research, there is a need to evaluate the mitigation measures that have been undertaken to meet the QIs, to design new interventions to further adhere to the QIs and for a broader, international perspective. Our practice testing shows that interventions should predominantly focus on the development of a structured, systematic approach for the organisation of PMAaH, proper documentation, policies on patient information and selection criteria for patients and locations. For the evaluation of these interventions, the presented QI set is a useful tool for single-centre and multicentre studies. Since the research is executed within the context of Dutch large teaching hospitals, an international perspective could be added by either having an international panel meeting and/or an additional practice testing in an international context. Besides, other hospital types, such as small regional hospitals or large academic hospital networks, could be involved for a validation of our QI set. Additionally, practice testing with home healthcare organisations and other organisations involved in the PMAaH process is important to provide a complete perspective on quality assurance of PMAaH processes. In the development and deployment of interventions based on this QI set, these organisations should also be actively involved. Lastly, since we have adjusted the RAM accordingly for the development of a structure and process QI set, we recommend that future research considers these adaptations. First, we have proposed a panel meeting coding scheme (online supplemental appendix D), which provided us with a more structured approach for extracting the results from the second part of the second phase. Note that we have not validated this coding scheme, which we propose as a future research direction. Second, we have adapted the third retrospective phase to fit to structural and procedural outcomes. This adaptation could add value to other RAM studies, but since our results show that there are relatively many changes after this third phase, we recommend that other studies include a second questionnaire to validate the suggestions from the third expert panel. This could increase the validity of the results. Furthermore, we believe that this practice testing elevates the RAM in these types of studies, which then elevates the impact in practical settings as well.

Data are available upon reasonable request.

The authors would like to thank the mProve working group medication@home for their initiative of, their support of and contribution to this research. We also would like to express our gratitude to Dr Alexia Schinagl for checking the final set on comprehensibility and proper use of the English language. We would also like to thank all experts who were included in any of the consensus round and/or practice testing for their valuable contribution.