Background

Person-centredness is at the centre of holistic healthcare and a core commitment of the World Health Organisation [1,2,3,4]. In order to deliver child-centred paediatric healthcare, it is essential to understand what is important to children and their families [5, 6]. The United Nations Convention on the Rights of the Child emphasises the importance of children being involved in matters that affect them [7]. Patient-reported information is central to improving care and quality of life, and evidence demonstrates that children can reliably self-report [6, 8]. However, their voices have not always been prioritised in clinical care or research [9].

Person-Centred Outcome Measures (PCOMs) are standardised questionnaires used to assess patient (and sometimes family) outcomes of healthcare [10,11,12]. They are usually self-completed by the patient, or proxy-reported when a patient is unable to self-report [10,11,12]. Research demonstrates that PCOM use can improve care quality and patient outcomes [13, 14], support conversations about care, initiate decision-making through shared language, and empower patients and families [11, 15, 16]. Whilst PCOM use has become common place and the benefits recognised in adult healthcare, there is limited understanding of the impacts, benefits, and implementation of PCOMs in paediatric services [11, 17, 18].

Additional complexities must be taken into consideration in use of PCOMs with children as opposed to adults, such as the need for child-centred language and their varying cognitive and developmental abilities [8]. Prior reviews have not incorporated the three aspects of implementation, service-focused and clinical outcomes [19] (e.g. on acceptability or improvements in Health Related Quality-of-Life (HRQoL)) and tend not to be theory driven limiting rigour and translatability [11, 20,21,22]. Theoretically-informed implementation strategies are needed to implement PCOMs into routine paediatric practice for the benefit of children, their families, and health care services (Including health and social care professionals, hereafter “professionals”) [23].

This systematic review aimed to identify and appraise the evidence for implementation of PCOMs into paediatric healthcare settings and develop a logic model to identify potential strategies for implementation and their causal mechanisms. The review objectives were 1) to identify determinants and strategies for implementing PCOMs; 2) to describe the mechanisms through which barriers and facilitators to implementation interact to enable or hinder implementation of PCOMs; 3) synthesise the findings through the development of a logic model; 4) to appraise the quality of the evidence.

Methods

This systematic review was conducted and reported in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [24]. The protocol for this review was registered on PROSPERO (International Prospective Register of Systematic Reviews; registration number CRD 42022330013).

Searches

CINAHL, Embase, Medline, and PsycInfo were searched to ensure articles across medical, nursing, and psychological disciplines were considered [8, 25, 26]. Google Scholar was searched for additional articles or grey literature and references cited in selected articles were also searched [27]. Databases were searched from 2009 to present (25 March 2022) as 2009 was the year the patient-reported outcome measure programme was introduced into the NHS in the UK [28] as well as a shift in thinking about and focus on outcome measurement in health internationally [29].

Search terms were informed by child-focused research [8, 26] and search strategies from adult PCOM and implementation research [12, 15, 25]. Related Medical Subject Headings were also used in conjunction with the keywords based on the following concepts: children AND outcome measures AND healthcare settings AND implementation. Full search strategies for each database can be found in the supplementary files [S1, S2, S3 and S4].

Study inclusion and exclusion criteria

Inclusion criteria

  • Population: children ≤ 18 years old. Studies which include both children and adults were included if the data about those 18 and under are reported separately, or if the population were professionals working with children, or their parents.

  • Intervention: Implementation or use of PCOMs or screening tools that are self-completed by a child in clinical care or proxy (parent/carer or professional) to improve care processes and/or outcomes.

  • Outcome: data relating to barriers and facilitators to healthcare implementation and/or sustained use of a measure.

  • Study types: Qualitative, case reports, quantitative (all experimental designs), mixed methods, service evaluations, quality improvement projects, audits. Systematic reviews were excluded but used for reference searching [27].

Exclusion criteria

  • Population: Studies including only people aged > 18 years where they are not professionals working with or parents/carers of children ≤ 18 years old.

  • Intervention: Studies where outcome measures are used to measure the effectiveness of an intervention or where measures are implemented into non-healthcare settings e.g., schools/social care

  • Outcomes: data relating to scores, psychometric properties, or reporting symptom prevalence only

  • Article type: Discussion/opinion articles, commentaries, editorials, letters, systematic reviews

Study selection

Articles identified in the search were imported to Covidence. HS screened titles and abstracts for eligibility; if there was not enough information to determine eligibility from initial screening, the full text article was screened. Full text articles were screened by HS and 10% were screened by a 2nd reviewer (DH). Discrepancies over eligibility of full text articles were discussed and resolved with a third reviewer (DB). Reason for exclusion of studies at the full text stage were recorded in a PRISMA flow chart [24].

Potential effect modifiers and reasons for heterogeneity

Heterogeneity in the data is anticipated due to the inclusion of paediatric healthcare settings globally and across multiple health conditions, therefore the barriers and facilitators identified may be context specific.

Study quality assessment

Study Quality Assessment was undertaken by HS. As multiple study types were included, several critical appraisal tools were used to assess the quality of studies of varying designs. The Critical Appraisal Skills Programme (CASP) tools [30] were used to assess study quality. Where there was not an appropriate CASP tool for the study design, the Joanna Briggs Institute (JBI) critical appraisal tools [31] were used. For mixed method studies, the Mixed Methods Appraisal Tool (MMAT) [32] was used. For quality improvement projects, the Quality Improvement Minimum Quality Criteria Set (QI-MQCS) [33] was used, and for non-randomised experimental studies of interventions, the Risk Of Bias In Non-randomized Studies – of Interventions (ROBINS-I) tool [34] was used. Articles were assessed against the items included in the checklists to develop understanding of the evidence rather than to exclude studies based on score. Study quality assessment results are presented in the results.

Data extraction strategy

Data was extracted by HS. Data extracted in Covidence included: authors, title, date, country, aim, design and methods, sample (including: conditions and age of child, proxy inclusion, inclusion/exclusion criteria, sample size), healthcare setting, outcome measure used, administration data (how it is delivered and by who), implementation data (facilitators and barriers [12]), and patient outcomes data. Data were extracted from both results and discussion sections to capture investigators’ observations regarding implementation of the measure. Where data were extracted from the discussion section of papers, this was noted.

Data synthesis and presentation

A narrative synthesis was conducted by HS to integrate qualitative and quantitative findings following the Guidance on the Conduct of Narrative Synthesis in Systematic reviews [35] with results discussed with RH, CES, DB. If disagreement occurred during these discussions, final adjudication (if needed) would be by RH. Preliminary synthesis involved tabulation, to develop initial descriptions of the studies and begin to identify patterns between studies. This was followed by a thematic analysis; deductively coding the extracted quantitative and qualitative data to the adapted-Consolidated Framework for Implementation Research (CFIR) constructs and sub-constructs [36, 37]. The adapted-CFIR comprises the original five domains from the CFIR with a sixth domain called ‘patient needs and resources’ [36, 37]. This gives person-centredness a greater focus to help ensure that patients’ needs are prioritised throughout all stages of the development, implementation, and evaluation of complex healthcare interventions [37]. This theory was selected as it is a well-established theory that has been evidenced to be effective for underpinning research and implementation of complex interventions in healthcare settings [23, 25, 40].

The effects of heterogeneity across studies were examined by comparing similarities and differences in outcomes across, study designs, settings, and populations to better understand the impact of context.

Logic model development

The adapted-CFIR supported the data analysis and subsequent development of a logic model using Smith et al.’s [38] Implementation Research Logic Model template by HS but presented and discussed with members of the research team (RH, CES, DB). The determinants of implementation in the template map to the adapted-CFIR constructs and sub-constructs [36, 37]. This allowed thematically coded data to be mapped directly into the logic model as either determinant barriers or facilitating strategies.

Results

Review statistics

Search yield

The search yielded N = 7401 articles from databases and a further n = 20 from citation searches. After duplicates were removed [n = 1789], n = 5632 records were title and abstract screened, and n = 5382 were excluded. Of the remining n = 250 records, n = 94 were conference abstracts and thus excluded. Following full text review [n = 156], n = 87 were excluded (reasons: no relevant outcomes reported [n = 36], adult population [n = 34], wrong study design [n = 8], wrong intervention [n = 6], wrong setting [n = 3]), with n = 69 retained for the analysis. Figure 1 below shows a PRIMSA Flow Diagram of the inclusion/exclusion process and Table 1 summaries the included studies.

Fig. 1
figure 1

PRISMA flow diagram. Adapted from Page et al. (2021) [41]

Full size image
Table 1 Summary of characteristics of included studies
Full size table

Study characteristics

Country

Of the n = 69 articles retained, n = 30 were conducted in the USA [42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71], n = 10 in the UK [72,73,74,75,76,77,78,79,80,81], n = 9 in The Netherlands [18, 82,83,84,85,86,87,88,89,90], n = 8 in Canada [91,92,93,94,95,96,97,98], n = 2 in Sweden [99, 100] and South Africa [101, 102], and n = 1 from each of Australia [103], Austria [104], Belgium [105], Germany [106], Iceland [107], Malawi [108], Norway [109], and North America (countries unspecified for the latter) [110].

Study design

With respect to study design, n = 17 studies used mixed methods [58, 65, 72, 75, 79, 82, 86, 89, 90, 95, 99, 100, 102, 103, 106, 107, 109, 110]; n = 15 studies used qualitative methods [46, 52, 55, 57, 60, 67, 70, 76, 78, 80, 91, 93, 94, 105, 108]. N = 12 studies were non-randomised experimental studies [44, 54, 61, 68, 69, 71, 77, 81, 85, 92, 96, 101], and n = 12 were quality improvement projects [42, 43, 45, 47,48,49, 51, 53, 62,63,64, 98]. There were n = 5 studies used cross-sectional designs [50, 59, 73, 74, 97], n = 4 case reports [18, 56, 66, 83], n = 3 cohort studies [84, 87, 104], and n = 1 study was part of a randomised control trial [88].

Setting

The most reported setting in articles was tertiary care hospital settings (n = 35) [18, 44, 46, 50,51,52, 54, 56, 58,59,60,61, 63, 69, 70, 79, 81, 82, 84,85,86,87,88,89,90,91, 94,95,96,97,98, 101, 103, 104, 106, 110], and n = 2 were conducted in both tertiary hospital and community settings [76, 105]. Of the n = 11 articles reporting on studies conducted in secondary care settings, these included: n = 8 mental health care settings [47, 62, 68, 72,73,74,75, 80], n = 2 speech and language clinics [92, 93], and n = 1 physiotherapy clinic [66]. A further n = 2 articles reported on studies conducted across secondary mental health care settings and the community [65, 77]. There were n = 12 articles reporting on studies conducted in primary care settings [42, 43, 45, 49, 53, 57, 64, 67, 71, 99, 100, 107, 109], and n = 2 articles reported on studies conducted across primary care and community settings [48, 55]. N = 2 articles reported on studies took place solely in the community [78, 108] and n = 1 took place across multiple settings, but did not specify details [83].

Population studied

Participants in included studies had a range of medical conditions, with many studies including children with multiple conditions. This included (using ICD-10 headings [111]): mental and behavioural conditions [42, 43, 45, 47,48,49, 54,55,56,57,58, 62, 64, 65, 67, 68, 71,72,73,74,75, 77, 78, 80, 92, 93, 99, 100, 102, 107, 109], cancer [18, 50, 59,60,61, 81, 83, 86, 87, 89, 90, 96, 103, 104], rheumatological conditions [50, 83, 84, 89, 90, 98, 106, 110], pain (chronic and acute) [44, 53, 63, 94, 95, 97, 110], endocrine conditions [46, 50, 51, 56, 82, 106], haematological conditions [50, 58, 61, 89, 90, 103], circulatory conditions (cardiac and pulmonological) [50, 56, 61, 69, 83], gastrointestinal conditions [50, 53, 89, 90], infectious diseases [83, 102, 108], respiratory conditions [56, 88, 106], neurological conditions [50, 95, 101], metabolic conditions [83, 89, 90], nephralgic conditions [56, 83], allergy/immunological conditions [56, 58], organ/stem-cell transplant [61, 91], other life-limiting/life-threatening conditions [76, 105], congenital conditions [70], ophthalmological conditions [79], and 4 studies included other chronic/unspecified conditions [52, 66, 85, 89, 90].

Measures studied and methods of completion

A range of generic and disease specific measures were used, including the Pediatric Quality-of-Life 4.0 (PedsQL 4.0), Psychological Assessment Tool 2.0 (PAT 2.0), and Quality Of Life in children and adolescents with DISABilities and their Families (DISABKIDS). There were n = 29 articles that reported on studies that included both child self-report and parent/caregiver-proxy report [18, 44, 47, 52, 56,57,58, 60, 61, 65, 66, 68, 72, 73, 81,82,83,84,85,86, 89,90,91, 95,96,97,98, 101, 105, 110]; n = 17 included child self-report only [45, 46, 51, 53, 54, 59, 62,63,64, 69, 79, 80, 88, 94, 104, 106, 108] and n = 18 included parent/caregiver-proxy report only [42, 43, 48, 49, 55, 67, 71, 77, 78, 87, 92, 93, 99, 100, 102, 103, 107, 109], usually because children were very young (< 6 years). There were n = 5 articles where this information was not reported [50, 70, 74,75,76].

Study quality assessment

Included articles varied in quality. As CASP do not recommend or include a scoring system [30], articles assessed with the CASP checklists were unable to be scored, however, 90% [n = 45] of articles that were able to be scored were assessed as of good to high quality (Table 2)—i.e., those that could be scored met > 80% of criteria or had low-moderate risk of bias.

Table 2 Summary of quality appraisal scores
Full size table

All qualitative studies (n = 15) reported clear aims and methodology addressing the research question. Weaknesses related to failures to discuss the relationship between researchers and participants, and lack of details of ethical considerations and recruitment strategies. The n = 1 randomised control trial reported a clear aim, but some methodological decisions were lacking, and participant demographics were not reported. The n = 3 cohort studies had clear, focused aims, and well detailed methods, but lacked detail about confounding variables and attrition.

Mixed methods studies (n = 17) met between 70–100% of the MMAT criteria indicating they were generally of high methodological quality. Methodologically weaker studies lacked detail of the sample and risk of non-response bias. Cross sectional studies (n = 5) were well-reported and on average met 98% of the JBI criteria. Case reports (n = 4) met on average 75% of the relevant JBI criteria. Non-randomised experimental studies of interventions (n = 12) assessed with the ROBINS-I tool were of low-moderate risk of bias; usually due to low adherence to the intervention, however these data were extracted as it pertained to the review aims. Quality improvement projects (n = 12) met on average 93% of the QI-MQCS criteria. Main areas of weakness were lack of reporting of patient-health related outcomes and data on the sustainability or scalability of the project.

Adapted-CFIR constructs

Table 3 details the adapted-CFIR domains and constructs extracted from the literature. The five most common subconstructs included complexity [n = 37], knowledge and beliefs about the intervention [n = 37], relative advantage [n = 31], patient needs and resources [n = 25], and available resources [n = 24]. Findings within each adapted-CFIR construct/sub-construct are presented below. Those constructs that were identified in less than three studies are not included in the narrative synthesis due to insufficient data [25]. Illustrative quotes are provided in Table 4 (reported as Q1, Q2, etc.).

Table 3 Factors identified using the Adapted-CFIR influencing PCOM implementation.
Full size table
Table 4 Illustrative quotes
Full size table

Intervention characteristics

Intervention source

Professional engagement in the development process of the measure [18] and the perceived security of the platform hosting electronic- or e-PCOMs were both factors that facilitated implementation [83]. However, low rates of parent completion were observed when newly implemented PCOMs were introduced to participants in the context of a research study rather than as new aspect of routine clinical care [Q1] [87, 99].

Evidence strength and quality

Presenting evidence to support PCOM use and perceptions of PCOMs as the ‘gold standard’ were key facilitating factors for implementation and frequent continuous use; training/education programmes that emphasised that PCOMs were research-evidenced, valid, and reliable supported this [70, 73, 92, 97]. Similarly, a significant barrier to implementation was professionals’ perceptions that there was insufficient evidence justifying PCOM use or supporting them as valid instruments [Q2] [50, 55].

Relative advantage

The use of PCOMs was perceived as advantageous, particularly from professionals’ perspectives. One study reported that 80% [n = 53] of parents found PCOM use provided added value over standard consultations [82]. Advantages included: improving communication, engagement, and decision-making with patients and families [Q3] [53, 55, 56, 61, 69, 79, 80, 91, 100, 103, 106, 108], enhancing quality of care and assessment [44, 46, 53, 56, 61, 62, 69, 70, 79, 80, 82, 85, 100, 105, 108], identifying concerns that would have remained unidentified in standard consultations [45, 46, 49, 53, 54, 56, 64, 79, 82, 100, 103, 105, 106], and increased referral rates and access to other services and treatment [45, 46, 48, 53, 56, 64, 103].

Three studies reported that professionals continued to use PCOMs after studies ended due to improved identification of patients’ unmet needs and due to the PCOMs having become integrated into routine practice [49, 64, 100]. Where professionals did not consider PCOMs beneficial, this was often due to them being perceived as bureaucratic exercises that did not elicit new information compared to standard consultations [Q4] [50, 55, 72, 93, 103].

Regarding e-PCOMs, there were mixed perspectives as to whether professionals felt that technology enhanced workflow and assessment compared to traditional paper-based PCOMs [49, 65]. However, there was a strong preference for accessing reports and scores from measures electronically [49, 50, 89, 90, 110], while inclusion of visual representations of progress, e.g. graphs tracking scores over time, was considered beneficial [44, 83].

Adaptability

Where PCOMs could be integrated into electronic systems or platforms this facilitated implementation [Q5] [57, 94, 106]. Correspondingly, this was identified as a barrier in studies where integration did not occur [42, 55, 72, 109]. Lack of cross-cultural validity of PCOMs and those not provided in service users’ language were identified as significant barriers [Q6] [52, 57, 78, 100, 108, 109].

Complexity

Ease of PCOM use facilitated implementation, which included professionals’ views of administering the measure, interpreting the score, and feeding back scores to patients and families [52, 57, 61,62,63, 65, 69, 71, 75, 78, 80, 82, 94, 100, 102, 106, 108, 110]. The importance of measures being child/user friendly was emphasised including ease of completion [Q7] [44, 49, 52, 58, 59, 61, 68, 70, 76, 78, 80, 86, 88,89,90, 94, 99, 100, 103, 104, 106, 108, 110], appropriate measure length [56, 58, 65], and language and a reading level understandable to children and parents completing the measures [Q8] [43, 59, 60, 67, 72, 76, 99]. However, the content of PCOMs (particularly items of a sensitive nature) was a barrier to implementation [Q9] [65, 80, 100].

Design quality and packaging

There was general preference for digital administration methods, such as tablets or computers [Q10] [58, 60, 98, 110]. Yu et al. [98] found 83% [n = 196] of parent/caregivers preferred the electronic version over paper or had no preference. Similarly, Stinson et al. [110] reported that only 16% [n = 77] of children preferred pen and paper as the method of administration. Children preferred using technology to complete measures (compared to professional administration), with 71% [n = 112] responding that they agreed or strongly agreed with the statement ‘I felt less embarrassed answering these questions on the computer than I would have with a clinician’ [68].

Cost

The two most significant costs associated with implementing PCOMs were monetary cost and time. Costs of implementing and maintaining e-PCOMs were discussed [43, 44, 98, 101, 108], although the initial cost of e-PCOMs could be offset over time (due to the recurring costs of paper-based measures [98]). There were some reports that measures were time-consuming to administer during appointments and concerns that this might detract from dedicated patient care [Q11] [43, 70, 108, 109]. However, in further study, providers reported that once PCOMs were implemented into routine care, an average of 16 min of time was saved per appointment [62]. This was particularly important as those time-savings were able to be redirected to improving patient care [62].

Outer setting

Cosmopolitanism

Multi-disciplinary, joined-up, inter-agency working was a significant factor in implementation, as there are often many agencies and services involved in the care of children [76]. Partnerships between settings facilitated the implementation and sustained use of measures and this was linked to peer pressure sub-construct [82, 106]. A lack of resource to address identified unmet need was a significant barrier to sustained use [Q12] [43, 50, 55, 57].

Peer pressure

Linking to cosmopolitanism and partnerships between settings, if other clinics they worked with were using specific PCOMs, this increased the motivation of professionals to also use them [70, 82, 106]; one study reported this was a motivating factor for 86.1% [n = 31] of paediatricians [82].

External policy and incentives

External recommendations, guidelines, or association endorsements were a motivating factor for settings to implement and use PCOMs in practice [80, 82, 92]. However, the source of the recommendation could potentially impact implementation [Q13] [108]. Lack of awareness of or disagreement with recommendations from professional associations was a barrier [55].

Inner setting

Structural characteristics

The main barriers regarding structural characteristics of organisations were related to organisational changes such as high staff turnover [43, 57, 75]. Age of professionals also impacted perceptions of using PCOMs in routine practice; in one study, older practitioners were more likely to be sceptical about the validity and evidence-base for using PCOMs [50].

Networks and communication

Multidisciplinary team communication was seen as a prerequisite to support use of PCOMs [91, 107]. Professionals recognised that using PCOMs supported information sharing between staff in a more systematic way, which improved care [Q14, Q15] [46, 61, 66, 103].

Tension for change

One potential barrier to the implementation of PCOMs was staff readiness and willingness to change current practice [55, 108]. However providing education and training to staff on the expected benefits of using PCOMs could change attitudes and willingness to change and thus facilitate implementation [47].

Compatibility

Perceived disruption to workflows was a potential barrier to implementation [Q16] [55, 60, 70, 93, 94, 98,99,100]. However, in practice, the introduction of PCOMs was generally not seen as disruptive and they became an integral aspect of routine care [Q17] [49, 60, 61, 63, 69, 82, 87, 108].

Relative priority

Shared recognition of the importance of using PCOMs, sometimes referred to as ‘buy-in’ or ‘ownership’ [60], was considered an important facilitating factor for implementation and use of PCOMs in practice [Q18] [60, 93]. Where professionals or patients did not perceive the benefit of PCOMs, this was a barrier to implementation [47, 91]. Education and training may have the potential to facilitate implementation where perceptions of PCOMs are a barrier as one educational intervention increased speech-language pathologists’ positive perceptions of outcome measurement from 49% of participants to 71% of participants [n = 46] [92].

Organisational incentives and rewards

Monetary incentives or rewards were a potential motivating factor for professionals to use PCOMs in practice, particularly in lower-middle income [112] countries [Q19] [82, 108]. Lack of reimbursement for administering tools was a barrier to implementation and continued use, particularly in countries with privatised insurance-based healthcare [Q20] [50, 55, 61, 67].

Leadership engagement

Commitment and support from leadership significantly contributed to the successful implementation of PCOMs into routine practice. High levels of support from leadership was more likely to enable successful implementation than when they had different priorities [Q21] [47, 57, 86, 107, 108], although this was not always the case [73].

Available resources

Lack of resources was a major barrier to implementation [43, 53, 55, 67, 72, 93] and continued funding was a necessity for sustainability [67, 108]. Lack of time to implement, administer, score, and record results of measures was a significant barrier, and often remained a barrier even when other barriers had been addressed [46, 53, 55, 57, 60, 63, 71, 80, 86, 93, 95, 109]. Inadequate staff numbers and high staff turnover was a barrier[Q22] [47, 57, 67], while recruiting additional staff to support measure implementation and use was a facilitator [46, 47, 57, 80, 93]. Challenges finding physical spaces for patients and families to complete measures in private was also an issue [Q23] [58, 60, 67, 82, 93, 108].

The technology requirements of e-PCOMs often created challenges, specifically in relation to internet access, access to and cost of devices [65, 67, 80, 94, 98, 101]. Conversely, paper-based measures had posed challenges pertaining to PCOM availability, stationery resources [42, 93, 94, 108], and the additional time required to enter results into electronic patient records [Q24] [65].

Access to knowledge and information

Lack of awareness and knowledge of both PCOMs and how to incorporate them into routine practice was a barrier to implementation and sustained use [50, 57, 64, 72], as was more knowledgeable staff not sharing their knowledge with those less knowledgeable [108]. Successful strategies to address issues of knowledge about using PCOMs in routine practice included reminders of when and to whom to administer PCOMs, through electronic health records or emails [Q25] [42, 48, 64, 93, 95]. Ongoing efforts to engage professionals through additional training, webinars, handbooks, and guidelines, improved sustainability [46, 47, 52, 57, 64, 72, 80, 85, 92, 100]. Explanations for children and families completing measures regarding what PCOMs are, their purpose, and how to complete them also facilitated use of PCOMs in practice [18].

Individual characteristics

Knowledge and beliefs about the intervention

When PCOMs were perceived positively, for example as validated tools that could support assessment and improve care outcomes, this acted as a facilitating factor [Q26] [44, 46, 49, 51, 53, 57, 58, 61, 68,69,70,71, 73, 74, 78, 79, 81, 84,85,86, 89,90,91,92, 94, 100, 102,103,104, 106,107,108]. Conversely, when PCOMS were perceived more negatively by either professionals or children and families (for example as time consuming), this acted as a barrier [Q27] [46, 50, 55, 67, 72, 74, 80, 91, 100, 108]. Educational strategies were often key to supporting implementation and use [72, 80, 85, 92, 100]. Additionally, if parents of children felt PCOMs were being used as tests, this could create unnecessary stress for families and act as a barrier [Q28] [76, 78].

Self-efficacy

Professional confidence [57, 82, 108] or lack of confidence [50, 57, 79, 93] in using PCOMs was a respective facilitator or barrier. Training and education to use PCOMs could increase self-efficacy and support implementation [46, 74, 108]. As professionals gained experience using PCOMs in practice, their self-efficacy increased [73, 99].

Individual identification with organization

Challenging relationships between professionals and management and a perceived lack of organisational commitment to the intervention were a barrier reported by one study [86]. Trusting relationships between professionals and families, and opportunities to work in partnership facilitated implementation [Q29] [60, 78].

Other personal attributes

Several personal traits were identified that could influence successful implementation and routine use in practice. Following through on actions was an issue for parents in terms of remembering to complete and return screening forms [42], and for professionals in terms of administering measures and discussing results with patients [42, 63, 83, 85, 93]. Professionals’ confidence, experience, and discipline all had the potential to act as barriers or facilitators [50, 55, 57, 73, 74]. Motivation of professionals and families was also important [Q30, Q31] [85, 89, 90]. For parents and patients particularly, motivation was often linked to the perceived added value of the measure for the consultation. Other personal attributes that could impact implementation included parental mental load and stage of treatment/diagnosis (which were often linked) [Q32] [87, 96] and how comfortable children felt talking to professionals [Q33] [91].

Process

Planning

Clearly defined responsibilities that have been collaboratively agreed with advanced notice are important for successful implementation of PCOMs [43, 46, 56, 57, 60, 63, 64, 70]. Absence of planning presents potential barriers to adherence [Q34] [57, 64, 93]. The way in which PCOMs are introduced and formally ratified by managers is also likely to have an impact on the success and uptake [Q35] [78, 86].

Formally Appointed Implementation Leads

Formally appointed implementation leaders or teams and support of site leadership were seen as essential components to adoption and uptake of newly implemented outcome measures [Q36] [56, 57, 65, 97, 107]. Lack thereof was noted as a significant barrier [43]. This sub-construct had significant cross-over with the Champions sub-construct, as the terms were sometimes used interchangeably.

Champions

Individual or team champions were seen as playing a key role in raising awareness of the interventions and promoting the use and value of PCOMs and supporting colleagues [Q37, Q38] [57, 60, 62, 65].

External change agents

External change agents who provide support, in terms of policy, advice, resources, or other forms of support to assist implementation, were seen as a facilitating factor [Q39] [65, 86, 107, 108].

Executing

As noted in some of the previously discussed sub-constructs, there were a number of logistical, resource, and education/information barriers which resulted in the intervention not being used according to plan; addressing these barriers was found to reduce these issues and increase adherence [42, 61, 63, 64, 86, 100]. However, patients and families forgetting or being unable to complete and return measures or completing the wrong measure was also an issue impacting implementation of PCOMs [Q40] [42, 52, 63, 100].

Patient needs and resources

Patient needs were better identified with the introduction of PCOMs into routine practice. PCOMs identified concerns of children and families that professionals perceived would not have been picked up in standard practice [Q41] [45, 46, 53, 54, 57, 60, 63, 64, 69, 77, 82, 87, 89,90,91,92, 102, 103, 105, 106, 108], with one study noting a 68% increase in identification [54]. PCOMs also increased referral rates though identifying unmet needs [45, 48, 53, 64]. Improvements in HRQoL scores were attributed to PCOMs supporting treatment decisions in one study, which reported 33% improvement in scores [56]. Increased focus on children, and better provision of individualised person-centred care were also noted [70, 76, 100, 105].

Evidence of effectiveness

Although several barriers to implementation were identified, numerous strategies from high quality research were able to successfully address barriers and support implementation of PCOMs into routine practice. In particular, training or educating professionals, children and families generally had a positive effect [42, 44, 47, 61, 63, 64, 73, 74, 80, 85, 92, 100] on the implementation of PCOMs, as did addressing logistical barriers [42, 63, 64]. Numerous studies also showed increased identification of concerns and referral rates after implementation of PCOMs [45, 46, 53, 54, 57, 60, 63, 64, 69, 77, 82, 87, 89, 90, 92, 102, 103, 105, 106, 108] which also acted as a facilitator for implementing PCOMs.

Logic model for implementing person-centred outcome measures in paediatric healthcare settings

The findings of this review have informed the development of a logic model (Fig. 2) which identifies determinants, strategies, and mechanisms for implementation from these barriers and facilitators. The logic model illustrates how the existing evidence for determinants of implementation can be used to develop strategies to achieve implementation, service, and patient/clinical outcomes. It also demonstrates the mechanisms though which these interconnected factors achieve outcomes.

Fig. 2
figure 2

Logic model for implementing PCOMs into paediatric healthcare settings. Adapted from Smith et al. (2020) [38]

Full size image

Discussion

This review has identified key barriers and facilitators to the implementation of PCOMs into paediatric healthcare practice using the adapted-CFIR. These findings informed the development of a logic model that can inform and support future development of context-specific implementation strategies for implementing PCOMs in different paediatric settings.

Relative advantage of PCOMs were echoed in the adult evidence base [13, 16] and some systematic reviews of measures used in specific paediatric settings [11, 20, 22], demonstrating benefits to decision-making, communication, identification of concerns, patient quality of life and referrals.

Existing evidence on implementation, emphasising the importance of PCOMs being evidence based, valid, and reliable [13, 14, 22] is reflected in the sub-constructs of intervention source and avidence quality. Øvretveit et al. [14] note the importance of measures being developed with the adult patients using them, to ensure suitability. A systematic review by Coombes et al. [8] suggests, in line with the findings from this review, that children generally prefer computerised measures and highlights the importance of measures being developmentally appropriate (relating to language used, recall period, and response formats) [8]. This further evidences the importance of involving key stakeholders in the development of PCOMs to support implementation and the mechanisms through which this occurs can be seen visually in the logic model.

The access to knowledge domain was intrinsically linked to relative advantage, intervention source, and evidence quality. In order for the intervention source, evidence quality, and relative advantage to act as facilitators, professionals, patients, and families must be supported to understand the reliability, validity and benefits of the PCOM [14, 18, 22].

Findings relating to resources, staffing, and leadership are consistent with literature from adult healthcare services, including the importance of integration within existing systems and workflows, staff willingness to change and ‘buy-in’, and leadership engagement and support [13, 14, 18].

Strengths and limitations

This systematic review provides a thorough, theory driven examination of the evidence for implementing PCOMs into paediatric healthcare settings. The adapted-CFIR supported the identification of facilitators and barriers to implementation, with only three sub-constructs for which there were no data identified. This supported the development of a comprehensive and theoreticallyinformed logic model. Given that 67% [n = 46] of the retained studies included child self-report measures, this review supports prioritisation of children’s voices in their care, and the centrality of person-centredness to quality care.

Of the sub-constructs for which no data was identified (trialability, individual stage of change, and reflecting and evaluating), it could be that these domains are not relevant to paediatric healthcare, or it could be due to limitations of the existing evidence base. These areas should be prioritised in future research.

Recommendations for practice

From the evidence synthesis and logic model development, several strategies for implementing PCOMs into paediatric healthcare settings have been identified. Education about the benefit of PCOMs is important to increase professional’s understanding of the importance and benefit of PCOMs to facilitate implementation [72, 80, 85, 92, 100]. Including key stakeholders in measure development helps to ensure the outcomes being measured are relevant and useful [8, 14, 18]. This further precipitates a sense of shared ownership with professionals, patients and families over the PCOM being implemented [60, 93]. The identification of context-specific factors (such as financial incentives as a facilitator in lower-middle income countries or reimbursement in privatised insurance-based healthcare systems) [50, 55, 61, 67, 82, 108, 112] further demonstrates the importance of professionals’ understanding the context in which implementation occurs.

Conclusions

To our knowledge, this is the first systematic review conducted into the implementation of PCOMs in paediatric healthcare that is not condition or setting specific. This review provides a comprehensive overview of the potential barriers to implementing and using PCOMs in paediatric healthcare, and the factors that can facilitate implementation and adherence.

This review has also demonstrated the suitability of the adapted-CFIR to theoretically inform implementation research in paediatric settings. The visual presentation of the logic model clearly demonstrates the interconnectedness of the numerous determinants of implementation. It also demonstrates the mechanisms through which implementation strategies can facilitate the implementation of PCOMs into paediatric healthcare settings to achieve improved outcomes for children and their families.

Using PCOMs in routine paediatric care is key to child-centred quality care. This review provides important evidence for how to implement PCOMs in practice in order to support better identification of patient needs. Future research should aim to assess the applicability and feasibility of this logic model in different settings to support implementation interventions, particularly in lower-middle income settings as much of the existing evidence come from higher income countries.