The study design was a cross-sectional survey, guided by the NPT and collecting quantitative and qualitative data using a web-based questionnaire. Findings are reported, when applicable, according to the STROBE checklist (45).

The study was conducted in the paediatric healthcare organisations in Region Västra Götaland, Sweden, comprising 26 clinics, and four rehabilitation clinics providing healthcare services for children with obesity. The organisations all cater to children with obesity and offer specialist health services. Region Västra Götaland is Sweden's second largest county council, providing healthcare services to approximately 1.7 million residents in western Sweden. One major city, Gothenburg, is located in the region, while the rest of the region comprises three smaller cities, medium-sized towns, and rural areas located in four regional areas (Table 1). In Gothenburg, PAP has already been introduced amongst healthcare professionals, through for example education, tutoring, networking, and PAP clinics supporting families whose children have been prescribed physical activity.

The inclusion criterium for participating in the survey was to be either staff or manager at a paediatric healthcare clinic or a rehabilitation clinic providing outsourced rehabilitation services for children with obesity, in Region Västra Götaland. No prior experience of working with PAP was required. Approximately 240 eligible participants were identified with the assistance of managers and administrative staff. The heads of departments approved the clinics' participation in the study; all 30 clinics accepted the invitation to participate.

All eligible participants were invited to answer a web-based questionnaire comprising four validated instruments measuring implementation outcomes. The questionnaire was distributed via e-mail during a three-week period in February and March 2021. To increase response rate and reduce the risk of non-response bias, three reminders at one-week intervals were sent.

In this study, the NPT was used to investigate and understand the collective work with PAP for children with obesity at the paediatric healthcare clinics. To assess the implementation process from the perspective of staff directly involved in the work of implementing PAP and their managers, the Normalization MeAsure Development tool (NoMAD) (42) was used. This instrument was specifically developed for implementation in healthcare contexts (42) and can be applied at any stage of an implementation process (46). It is adaptable to different interventions and settings, and can be combined with other measurements focusing on other dimensions of implementation (42).

The NoMAD instrument consists of 23 items, of which three general questions are indicators of normalisation answered on 11-point Likert-type scales ranging from “still feels very new” to “feels completely familiar” for item 1 and from “not at all” to “completely” for items 2 and 3 (46). Twenty items target the four core NPT domains: (1) coherence, that is the “sense-making” work people do to initiate a new intervention or practice; (2) cognitive participation, described as the relational work around the practice; (3) collective action, the work to perform/operationalise the practice; and (4) reflexive monitoring, the appraisal work to understand the new practice (42). Each item has two options, A and B. Option A is answered on 5-point Likert-type scales with response options “strongly agree”, “agree”, “neither agree nor disagree”, “disagree”, and “strongly disagree”. Option B is applicable only to those who found no relevance in answering option A (46).

In this study the validated Swedish version S-NoMAD (47) was used. The word “intervention” was replaced by “PAP for children with obesity” or just “PAP”. Although the instrument was developed for healthcare professionals directly involved in the intervention, we wanted to address the perspectives of both staff and managers. Some items were slightly modified by adding wording addressing managers, e.g., “Do you feel PAP is currently a normal part of your work/area of responsibility?” and “Management supports/I as a manager support PAP for children with obesity”.

To supplement NoMAD, the implementation determinants acceptability, appropriateness, and feasibility of implementing PAP in paediatric health care were assessed, adding the staffs’ and managers' perceptions and attitudes towards PAP. Acceptability is defined as the perception amongst stakeholders that a given treatment, service, practice, or innovation is agreeable, palatable, or satisfactory (44). Appropriateness is the perceived fit, relevance, or compatibility of the innovation or evidence-based practice for a given practice setting, provider, or consumer; and/or perceived fit of the innovation to address a particular issue or problem. Feasibility is the extent to which a new treatment, or an innovation, can be successfully used or carried out within a given agency or setting (44). These outcomes were measured with the Acceptability of Intervention Measure (AIM), Intervention Appropriateness Measure (IAM), and Feasibility of Intervention Measure (FIM) (48). All three are validated instruments with the purpose of assessing the fit and match of a practice or intervention to a given context, targeting different criteria (48). The measures comprise four items each, answered on 5-point ordinal scales with response options “strongly agree”, “agree”, “neither agree nor disagree”, “disagree”, and “strongly disagree”. The instruments were translated and cross-culturally adapted into Swedish, adapted to children with obesity, and validated (49).

In addition, we collected demographic data and data on PAP experience and prescribing frequency. Two open-ended questions explored barriers and facilitators, in which the respondents were given the opportunity to describe their own experiences and thoughts regarding determinants for implementing PAP.

The quantitative variables are presented descriptively using frequencies and percentages and medians and interquartile ranges. To facilitate future comparisons with other studies, means and standard deviations (SD) are also presented. The respondents' practice location was categorised into Gothenburg and other regional areas of the Region Västra Götaland (Table 1). Work experience in the organisation was categorised into (1), <2 years; (2), 2–5 years; (3), 6–10 years; and (4), >10 years. Professions were categorised into six groups: (1), nurse; (2), physician; (3), dietician; (4), physiotherapist; (5), manager and (6), other. Number of years working with/taking decisions about PAP was categorised into (1), <3 years; (2), 3–5 years; and (3), >5 years or longer. Missing data analyses were performed using chi-square tests to examine proportion of managers and practice location amongst non-responders vs. responders.

To facilitate presentation and interpretation, response categories were merged into fewer categories. Responses to the three general items were coded as: 0–4 = not familiar and 5–10 = familiar for item 1; 0–4 = not a normal part of work and 5–10 = a normal part of work for item 2; and 0–4 = it will not become a normal part of work and 5–10 = it will become a normal part of work for item 3. For the NoMAD items, the disagree/strongly disagree response categories were merged into disagree, and the strongly agree/agree response categories were merged into agree. Item 3.2 was reverse-scored due to its negative wording. One item (2.2) was not analysed since it was accidentally removed from the questionnaire. For the AIM/IAM/FIM suite of instruments, the response categories completely disagree/disagree were merged into disagree and agree/completely agree were merged into agree.

Comparative analyses of participants from the Gothenburg clinics in which PAP has already been introduced vs. clinics in the rest of the region were performed using chi-square tests for the dichotomised general questions. Because the assumptions of the chi-square tests were not met for the NOMAD and AIM/IAM/FIM items, we performed Mann–Whitney U tests using the original 5-point scales. Differences between years of work experience in the organisation and between professions in all variables were performed using Kruskal–Wallis tests with pairwise comparisons, applying Bonferroni correction. For variables where there were significant differences in the main Kruskal-Wallis test, we only report significant differences in the pairwise comparisons. Because age and work experience correlated, no comparisons were made between age groups. A p-value of ≤0.05 was considered statistically significant. All cases for which all items in at least one instrument were completed, were included in the analyses.

Internal consistency of the NoMAD items was acceptable for coherence (Cronbach's α = 0.748), and questionable for cognitive participation (α = 0.600), collective action (α = 0.638), and reflexive monitoring (α = 0.687). For the the AIM/IAM/FIM measures, internal consistency was excellent for acceptability and appropriateness (Cronbach's α = 0.924 and 0.943, respectively) and good for feasibility (α = 0.892). Quantitative data were analysed using IBM SPSS, version 28 (IBM Corp, Armonk, NY).

The barriers and facilitators described in free text answers to the open-ended questions were coded and sorted into categories corresponding to the NPT domains. This was done in an iterative process by the first author together with two physiotherapist colleagues with experience of working with PAP for children.

A majority of the respondents (81.1%) reported being familiar with PAP (Table 3). A higher proportion of respondents in the Gothenburg area reported being familiar with PAP than those in regional areas (90.0% vs. 70.6%, χ² = 6.772, p = 0.009). Physiotherapists reported familiarity with PAP to a greater extent than “other” professions (Mdn 9 vs. Mdn 5, U = 55, p = 0.025). Fifty-six percent described PAP as currently being a normal part of their work; a higher proportion of respondents from the Gothenburg area reported this than those in the regional areas (70.0% vs. 40.7%, χ² = 9.882, p = 0.002). Nurses reported feeling PAP was a normal part of their work to a greater extent than “other” professions (Mdn 6.5 vs. Mdn 0.5, U = 38, p < 0.001), as did physiotherapists (Mdn 9 vs. Mdn 0.5, U = 53, p < 0.001). A majority (82.0%) reported believing that PAP will become a normal part of their work. Respondents with >10 years of work experience in the organisation reported this belief to a greater extent than those with 2–5 years' experience (Mdn 9 vs. Mdn 6, U = 23, p = 0.012).

Most respondents (67.9%) agreed that they could distinguish between PAP and their usual ways of working, and 56.1% reported that they have a shared understanding of its purpose (Table 4). Respondents in the Gothenburg area agreed to a greater extent than those in regional areas to having a shared understanding of PAP (Mdn 4 vs. Mdn 3, U = 1170, p = 0.005) and of how the intervention affects the nature of their work (Mdn 4 vs. Mdn 3, U = 1104, p = 0.017) (Table 5). About three quarters of the respondents (76.6%) agreed on the potential value of PAP. No differences were seen related to work experience in the organisation or profession in this domain. Option B responses were selected by 3–8 respondents (2.6% to 6.8%).

A barrier for using PAP described in the open-ended questions was the respondents' experiences of not knowing the PAP intervention well enough and working with single components alone, particularly the written prescription for physical activity. The opposite, a comprehension of the PAP intervention and considering and including all of its components, was described as a facilitator. Statements like “I consider it important that PAP is well supported by a good assessment so it will be at the right level, for example goal setting, activity, duration, and that the patient is motivated. If not, then it might just be ‘another piece of paper’ for the individual.” were typical.

Almost half (47.2%) agreed that there are key people who drive PAP forward and get others involved. Respondents in Gothenburg agreed to a greater extent than those in regional areas that there are key people driving PAP forward and involving others (Mdn 4 vs. Mdn 3, U = 1000, p = 0.003). Most reported being open to working with colleagues in new ways to use PAP (94.5%) and agreed to continuing to support PAP (85.7%). No differences were seen related to work experience or profession in this domain. Option B responses were selected by 1–8 respondents (0.9% to 6.8%).

A reported barrier in this domain for using PAP was the absence of physiotherapists at the clinics and the perceived uncoordinated pathways to healthcare units offering PAP support. Facilitators for using PAP were colleagues being supportive of PAP and successful healthcare collaboration. Statements like “In my clinic we have divided the tasks between us a little. However, I could prescribe PAP more often, but mostly it's done by my colleague who is a nurse.” were reported.

Over half of the respondents (57.8%) agreed they can easily integrate PAP into their existing work and only 1.8% agreed that PAP disrupts working relationships. A majority (77.2%) reported having confidence in their colleagues’ ability to use PAP. Over half (56%) agreed that work is assigned to those with skills appropriate to PAP. One fourth (26%) agreed that sufficient training is provided to enable staff and managers to implement PAP. Respondents in Gothenburg agreed to a greater extent than those in regional areas that work is assigned to those with skills appropriate to PAP (Mdn 4 vs. Mdn 3, U = 1116, p = 0.032), that sufficient training to implement PAP is provided (Mdn 3 vs. Mdn 2, U = 704, p < 0.001), and that sufficient resources to support PAP are available (Mdn 3 vs. Mdn 3, U = 1121, p = 0.029). No differences were seen related to work experience or profession. One fourth (26.9%) reported that sufficient resources are available to support PAP and half (51.1%) agreed that management adequately supports PAP. Option B responses were selected by 1–22 respondents (0.9% to 19.5%).

Barriers from the open-ended questions were inadequate education and insufficient time to use PAP. Statements like “I would like to learn more about PAP, but I have too many duties to have time to plunge into it. It's not my most prioritised task, instead it's something I do on the side, a few times a month” were typical. Facilitators were staff taking on the role of using PAP and having more time with patients when delivering PAP.

Thirty-seven percent reported being aware of reports about the effects of PAP. Managers agreed to a higher extent than “other” professions that they were aware of reports (Mdn 4 vs. Mdn 3, U = 45, p = 0.013) and respondents with more than 10 years of work experience in the organisation agreed to a higher extent than those with 6–10 years of experience that they were aware of reports (Mdn 4 vs. Mdn 2, U = 30, p = 0.007). Sixty percent agreed that PAP is worthwhile and 48.5% valued the effects PAP has had on their work. Respondents in the Gothenburg area agreed to a greater extent than those in regional areas that they valued the effects (Mdn 4 vs. Mdn 3, U = 890, p = 0.011). The respondents agreed that feedback about PAP can be used to improve it in the future (81.2%). No differences were seen related to work experience. Option B responses were selected by 2–22 respondents (1.7% to 12.1%).

A reported barrier for using PAP was the lack of research on PAP for children. Statements like “I’d like to see randomised studies that are large enough to show the effectiveness of PAP if I am to become positive about the intervention” are illustrative. The opportunity to provide discounted activities was reported as an important facilitator.

Most respondents stated that PAP meets with their approval (85.6%), is appealing (85.6%), and that they like (84.0%) and welcome (83.2%) PAP (Table 6). Respondents in the Gothenburg area agreed to a greater extent than those in regional areas that PAP meets their approval (Mdn 5 vs. Mdn 4, U = 1528, p = 0.022). Respondents with more than 10 years of work experience agreed to a higher extent than those with 2–5 years of experience that they welcome working with PAP (Mdn 5 vs. Mdn 4, U = 23, p = 0.019). No differences were found by profession.

Most agreed that PAP seems fitting (81.6%), suitable (83.2%), applicable (80%), and like a good match (78.4%) for children with obesity (Table 6). No differences were seen by practice location, profession, or years of work experience.

Most respondents reported PAP being implementable (78.4%), possible (88%), doable (82.4%), and easy to use (71.2%) (Table 6). No differences were found by practice location, profession, or years of work experience.

A main strength of the study is the use of a theory-based framework and instrument to assess and categorise the factors that might influence implementation of PAP in the paediatric context. A particular strength in using the NPT is its focus on the implementation work healthcare professionals actually do, rather than their cognitions, e.g., beliefs and attitudes. Another strength is our use of validated instruments, which are also pragmatic and easy to use. The NoMAD was particularly helpful in pointing out problems that can be addressed when implementing PAP for children with obesity, enabling improvements related to collective action and reflexive monitoring. Assessing the dual perspective of practitioners and managers also strengthens the findings. Supplementing NoMAD with the AIM, IAM, and FIM instruments to assess important implementation determinants provided a comprehensive overview of aspects necessary to address in a future implementation of PAP for children with obesity. Several efforts were made to reduce bias. Sampling bias was minimised since the survey was distributed to all staff and managers at all paediatric clinics in the study population. We attempted to reduce non-response bias by sending several reminders to answer the questionnaire.

There were some limitations to the study. The intention to capture multiple perspectives meant that not all participants had practical experience of PAP, making several questions irrelevant for some respondents and likely contributing to both unit-level and item-level missing data. The use of self-reported data entails a risk for both self-selection bias and social desirability bias. We did not perform sensitivity analyses, but believe our analyses are robust enough with the used tests. We did not investigate gender, because a vast majority of both practitioners and managers in the population studied are women. The low alpha values for some of the NoMAD items indicate low internal consistency, which might have affected the results of the statistical analyses.

There is an obvious need for research on effectiveness of physical activity promoting interventions for childhood obesity, as well as implementation process and outcome evaluations of such interventions. To improve the understanding of barriers and facilitators for using PAP, further research is needed from the perspective of staff and managers, as well as that of the children and their parents.

Our study can provide helpful information to develop support structures for PAP work, streamline the use of the intervention, and inform future implementation strategies. The broad inclusion criteria of the study, including all professions and managers involved in paediatric health care, and the study setting – Region Västra Götaland which is Sweden's second largest county council – enhances generalisability of our findings to other paediatric populations and to other regions in Sweden, and possibly also to other countries with similar paediatric healthcare systems.