Practice variation amongst preventive child healthcare professionals in the prevention of child maltreatment in the Netherlands: Qualitative and quantitative data

This article provides both qualitative and quantitative data on practice variation amongst preventive child healthcare professionals in the prevention of child maltreatment in the Netherlands. Qualitative data consist of topics identified during interviews with 11 experts (with quotes), resulting in an online survey. The quantitative data are survey responses from 1104 doctors and nurses working in 29 preventive child healthcare organizations. Additionally, the interview topic list, the qualitative data analysis methodology, the survey (in English and Dutch) and anonymized raw survey data (http://hdl.handle.net/10411/5LJOGH) are provided as well. This data-in-brief article accompanies the paper “Variation in prevention of child maltreatment by Dutch child healthcare professionals” by Simeon Visscher and Henk van Stel [1].


a b s t r a c t
This article provides both qualitative and quantitative data on practice variation amongst preventive child healthcare professionals in the prevention of child maltreatment in the Netherlands. Qualitative data consist of topics identified during interviews with 11 experts (with quotes), resulting in an online survey. The quantitative data are survey responses from 1104 doctors and nurses working in 29 preventive child healthcare organizations. Additionally, the interview topic list, the qualitative data analysis methodology, the survey (in English and Dutch) and anonymized raw survey data (http://hdl.handle.net/10411/5LJOGH) are provided as well. This data-in-brief article accompanies the paper "Variation in prevention of child maltreatment by Dutch child healthcare professionals" by Simeon Visscher and Henk van Stel [1]. &

Value of the data
Lists 38 topics that interviewed experts believe are vital in the prevention of child maltreatment, yet are not commonplace at the moment Provides response distributions for each questionnaire item separately Can be used to prioritize specific quality improvement efforts May inspire other practice variation studies, because the data provides a frame of reference and extra insight into our methods enhances reproducibility of each step during the process Provides the English and Dutch versions of the survey on prevention of child maltreatment by preventive child healthcare professionals

Data
First, we present the interview methodology, including the selection and the characteristics of the experts we interviewed. Second, we describe the qualitative methods used to analyse the interviews. Third, topics emerging from the qualitative analysis are listed, which the experts considered important in the prevention of child maltreatment, yet, of which they expected a large amount of practice variation would be present. Fourth, we present the questionnaire we used to verify this hypothesized practice variation. For each multiple choice question the response distribution is displayed. For some questions additional data are provided.

Interview methodology
Individual interviews were conducted, for which the subjects were recruited using purposive sampling [2]. The objective was to select individuals who should be qualified to define optimal care in this area. The interview was based on a topic guide, which was augmented with new topics after each interview. The initial topic guide was based on the guideline and the current literature [3,4]. After informing the participant of matters such as anonymity and study objective, the interview started with a broad open-ended question [5]. As the subject responded, neutral probing questions were asked to identify specific topics, and to determine the relevance of each topic for inclusion in the questionnaire. Subsequently, topics were introduced, which had not yet been mentioned spontaneously by the interview subject. Once the topic guide became too lengthy to cover each topic in every interview, topics that were satisfactorily discussed in the foregoing interviews were omitted. New interviews were conducted until data saturation occurred (i.e. no new topics or viewpoints came up). Several techniques were used to improve the quality and validity of the data, including audio recording and member checking of the transcripts [5,6] (Tables 1-5).

Interview participants
Eleven experts were interviewed (age 30-68 yr), including 7 physicians, 3 nurses and 1 health scientist. Several participants were highly experienced because they either worked at child protective services (n ¼1), taught prevention of child maltreatment (n ¼3), or occupied the position of "designated expert" at their own child healthcare organization (n¼4). The latter involves being available for consultation by colleagues, which is standard procedure according to the national guidelines, every time a child healthcare professional suspects maltreatment. Other participants (co)authored a book on child maltreatment (n ¼2), or contributed to the current guideline (n¼ 4) and/or its upcoming revision (n ¼4). Professionals active in both age categories, 0-4 and 4-19 yr participated.

Qualitative analysis
The qualitative data-analysis consisted of organizing and interpreting the data. The first part was done by thematic analysis: assigning codes to chunks of transcribed text relating to a particular topic [7] (using NVivo 10 [8]). Once all substantive text was coded, an overview was generated for each code, containing the passages related to the topic. Subsequently, these overviews were interpreted and an overall summary was made. Whilst summarizing, a five-point scale was used to describe the importance of a topic, and its expected amount of variation. Based on the summary, an overall score was calculated, which represented the overall relevance for inclusion in the questionnaire. A detailed description about the interpretation process can be found below. Finally, the topics were classified into categories. Topics

Topic Description
Openness Openly communicating suspicions towards the parent, as opposed to secretively collect evidence against him/ her. Child maltreatment is almost always the cause of pedagogical incompetence, rather than malice. The first step should almost always be to communicate your observations with the parents. "You shouldn't have a hidden agenda." "You have to dare to openly share your worries."

Open-mindedness
Communicating to a client that you will not judge them, so that they can tell you anything. "This part also has to do with norms and values. (…) for instance, a mother who kept her child on a leash outside, because her child would otherwise run away. [You mean like people normally do with their dogs?] Yes. And there were many other things as well. But yes, I found that abnormal. And that is not per se directly child maltreatment, but it was a whole aspect, she also… (…) I think that's close to child maltreatment. Or maybe it is child maltreatment, systematically, because the child also had all kinds of behavioural problems. While she thought: yes but we're doing so well. I don't see the problem and uhh… I'm sharing this with you now, but you are judging me for it. Well, from now on I won't share anything anymore. (…) So you need to be able to empathize with why the parents do things. Projecting yourself into their reference frame." "Especially things like sexual abuse, are veeery sensitive topics. It is veeery hard to make that discussable, to then just think like, what could this be? Because, and to then, not have a verdict on that, but to openly discuss it with the parent…" Conversational skills Being skilled at communicating with children in specific, and being able to interpret the parent-childinteraction.
"If a child says that daddy was very angry, and you say, well, did smoke come out of daddy's ears, or that kind of stuff, and the child goes on with it, and says yes, there came so much smoke out of his ears and then he exploded! And a parent is sitting next to the child and thinks well let him babble… (…) Parents can of course brilliantly tell how well they are doing, whereas children can also say they're okay, but I don't like this, or I don't like that, and… (…) children are of course part of the story." "Being able to pick up signals about does a child feel safe or do they feel unsafe? A child who has never seen me, walks into my consultation room and, just like that climbs on my lap, I'm worried about such a child. That kind of signals." Diplomacy Being skilled at using diplomatic/tactful language. Prioritizing in situation with multiple problems that need to be solved. Being able to negotiate with parents.
Being able to bring a message across, without ruining the relationship. "If say to a Moroccan father like, in Holland it's forbidden to beat a child, in common, so, the teacher isn't allowed to do it, I don't offend him, but then he knows it. And he will act like he already knew that, but in the meantime he thinks hmm…" Domain: Medical expertise Topic Description Knowing all risk factors and symptoms Knowing all risk factors and symptoms of child maltreatment. After openly communicating suspicions with parents, this was regarded the most important topic of all. "So you have to know the signals and all, the hundreds of signals and symptoms very well, of child abuse, maltreatment and sexual abuse." "Sensitivity, but also knowledge. You also have to know what all the possibilities are, when it comes to risk factors and signals." "Then I wonder is he appropriately dressed for the season? You know, that kind of stuff." "When it comes to babies, half a year of age, you can always easily see if the attachment is secure. If I examine that child, and I am standing above that child, then I am a threat. And what are you then supposed to see: a child of half a year old looks at the mother. And s/he reads the face of the mother, the mother smiles, then it is okay. That is a sign for me that the child is safely attached to his mother. A professional can also organize events for his own colleagues, perhaps in collaboration with child protective services (see also: frequent education). Domain: Involvement Topic Description Overall involvement related to child maltreatment Being involved in the prevention of child maltreatment on micro-, meso-and macro level: on a case level, but also involvement on a neighbourhood level (e.g. "You know easy, that it's is not, but that you say like, you know it occurred to me that we see a lot of teenage mothers in this neighbourhood. How can we make sure that these teenage mothers do well?"), on the municipality level (e.g. influencing the policy of the municipality) and on larger levels (e.g. teaching about signals of child maltreatment at a regional educational institute for teachers, or being politically active). "There are people who really, how shall I put it, well who read about it and that's it, but you have to integrate it in your, in your whole professional profile, it has to be intermingled. It shouldn't be yet another subject. (…) It consists of being able to, daring and willing to see child maltreatment." Consultation of Child Protective Services Always contacting CPS when child maltreatment is suspected. Many professionals think CPS is only meant for filing a report of child maltreatment, but this is not the case. They usually offer invaluable guidance on how to communicate with parents and which subsequent steps should be taken. In the Netherlands it is mandatory that professionals who work with children, get in touch with CPS when they have suspicions of child maltreatment. Since the interviewed experts estimate that 10-20% of the professional's workweek should primarily be focused on preventive activities related to child maltreatment, the number of consultations with CPS should be substantial.

Consultation of internal child maltreatment expert
In the Netherlands every preventive child health care organization is obligated to appoint an internal expert on child maltreatment. The national guideline on the prevention of child maltreatment recommends that this expert is contacted each time a professional suspects child maltreatment. The interviewed experts believe this is a very important recommendation, which should always be observed. Frequent education on the topic (of good quality) Regularly attending a course on the prevention of child maltreatment, which also has to be of good quality.
"Well, what I always see as internal expert, is that if we have given a refresher course again and it's just put back on the map again, that the first two months thereafter we get phone calls more often, from people who say 'I want to discuss this case with you, because I don't know what to do with it.' So they are doing it again, so in one way or another, you have to continuously bring it back to their attention, because things simply have a tendency to fade away. That works for sure." Eagerness to improve oneself When it comes to prevention of child maltreatment, professionals have to be eager to raise the bar. Domain: Improvement opportunities Topic Description Fear of being wrong Fear of being wrong or conflicting harm to the child. This fear was expected to be a very important impeding factor.

Fear of reaction
Fear of how the parent will react. This fear was expected to be a very important impeding factor. Fear of damaging the relationship Fear that parents will stay away after suspicions are shared with the parents. This fear was expected to be a very important impeding factor. Fear of suffering rebuke Fear of being rebuked by their manager, because the parents file a complaint, or because their documentation falls short. This fear was expected to be a very important impeding factor. Fear of breaching medical confidence Fear of breaching medical confidence and/or being punished by the board of medical examiners. This fear was expected to be an important impeding factor. Lack of time for acting upon suspicions When a professional signals child maltreatment, they may not always act upon it due to lack of time. Child healthcare physicians do not feel respected by other medical specialties, leading to low self-efficacy. Some interview participants believed this negatively influences their assertiveness. Notwithstanding, they do see an improving trend concerning this issue. "So I think the Calimero-effect, insecurity, not respecting their own specialty. (…) It matches their place on the medical ladder. You have to fight for that. (…) They should fight, but instead they flight or freeze.", "We see that younger physicians are stronger, they already have some history in the hospital, during the educations, that has strengthened them. But we do see a relatively large groups that finds it difficult.(…) We need the innovators who will say, yes, it is also an opportunity to change things for the better." Items that were left out of the questionnaire Topic Description

Home visits
A large number of standard home visits (as determined by the organization), as well as a lower threshold to conduct a home visit on indication, is expected to enhance the prevention of child maltreatment.

Number of standard contact moments
The number of standard contact moments can vary a (little) bit between organizations. Most organizations don't have a contact moment between the age of 5 and the age of 11 years. A higher number of contact moments is expected to aid the prevention of child maltreatment.  Table 2 Interview design.

Schedule
Introduction (see "Techniques to improve quality of the interview data") Full-breadth open-ended question [5], enhanced by: Neutral probing questions Probing questions to ensure coverage of all aspects Skills, knowledge, methods and improvement opportunities Importance, as well as expected amount of practice variation Clarifying questions and summaries Introduction of items from the topic guide that were not mentioned spontaneously Question about coverage of the interview (i.e. "Where all major topics discussed?") After each interview: topic guide update (based on notes from during the interview) Initial topic guide

Practices
Adherence to the child healthcare guideline for secondary prevention of child maltreatment Adherence to national reporting laws Self-improvement (e.g. studying, reflection) Activities to promote acquaintance with other agencies No experience in scientific interviewing Table 4 Coding methods.

Coding [7]
The researcher was already familiar with the data Design of the interview schedule and its topic guide Administration of the interview Transcription of the interview Node list The topic guide served as an initial list of nodes. If a passage was eligible for multiple nodes, each node was first precisely defined. If the problem persisted, the passage was multicoded.
New codes were created as additional concepts were identified. Existing nodes related to the new nodes were fully recoded. Table 5 Scoring methods.

Scoring
Principles Goal of the scoring system is to prioritize topics relative to each other, for selecting items for the questionnaire Focus on relative credibility, instead of absolute credibility Overall impression of the researcher is conclusive Primary focus on the literal meaning of the text and triangulation Secondary focus on the context of the passage Tertiary focus: quantitative analyses performed in NVivo (word count, passage count, interview count, interview coverage (i.e. word count divided by total interview word count) and overlap with other nodes) Triangulation Opinion of one participant: devaluation of literal meaning of the text Opinion of two or more participants: inclination towards literal meaning of the text Opinion of three or more participants: acceptance of literal meaning of the text Triangulation can be overruled by compelling arguments, such as: Scientific evidence Legitimate examples and a reasonable explanation of underlying mechanisms Participant is especially accomplished in the particular area The participant finds it extremely important, at least supported by a high interview coverage Spontaneously shared information has more value than answers to specific questions Importance Definition: How important this topic is for the efficacy of primary or secondary prevention of child maltreatment. þ þ Very important þ Important 7 It helps -Not important, but it probably helps -Not important, and it probably does not help either Expected amount of variation Definition: The amount of improvement that could probably be made through change of policy.* This measure is a composite score of how many professionals could improve in this area, and how much they could improve in this area.
* ¼ The influence of policy was estimated by the researcher þ þ most professionals could improve a lot þ it is not uncommon that professionals could improve a lot OR most could improve moderately 7 it is not uncommon that professionals could improve moderately OR most could improve a little -it is not uncommon that professionals could improve a little in this area only exceptional cases could improve in this area OR the influence of policy is small Expected amount of response bias (conjecture of researcher) Definition: The amount of bias the researcher (subjectively) expects, considering the fact that only a few brief questions can be used, and taking into account social desirability, probability of overestimation, subjectivity and complexity.
þ þ no notable consequences are expected þ a meaningful estimation of the effect size is expected to be possible 7 a meaningful estimation of the effect direction is expected to be possible -a meaningful estimation of the effect direction may be possible, dependant on the average tendency towards response bias this item should not be measured by a short question, because interpretation of the results will not be possible. Overall relevance score þ þ: Very relevant -(all items Z 7 ) AND ( The personality traits of being more responsible and/or being more compassionate were expected to strongly impact child maltreatment prevention. Nonetheless, they were left out of the questionnaire because we do not think they can be measured reliably using selfassessment. Furthermore, they are less relevant for this study than the aforementioned topics, because it is hard to influence these areas, making them less suitable for quality improvement efforts.

Questionnaire -English version including responses
Page " N ¼68 (20%) Other finished courses (incl. non-CHC related), namely: (textbox) Comments: The "currently attending" response option was not analysed, because it was only included to avoid response bias from CPs who currently attend a course or are nearly finished with a course. This question actually addresses two topics: profession and specialisation. These were merged into a single question, to distract the respondent from filling in their profession. Since competition between the two types of professionals cannot be excluded, this decision may have reduced social desirability bias (i.e. "faking good") [9]. 4. Please fill in the following details: Mean "(Name of city) (Neighbourhood: (name of neighbourhood))" b.
"(Names of three small villages)" N sufficient detail ¼ 877 * in the questionnaire real names of places were used.

Page 3
5. How often does your team organize an activity for parents and/or partner organizations in which you either tell them when to contact you, or in which you give them advice? (E.g. a morning for teachers about signals of CM; or for parents about child rearing.) My team does this (Attention! There are two boxes. Choose for example "1 time" in the first box, and "per year" in the second box) Dropdown menu The question is asked in a relatively factual way, which probably reduced self-overestimation. All six partner organizations were described by the interview participants as "important to know someone from, in person", and except for the emergency department, this was also consistently described as feasible. 7. Which proportion of the professionals below contacts you very easily, if they suspect child maltreatment (estimation): For example, suppose 3 out of 6 GPs in your working area very easily contacts someone of your team: select the bullet in the middle.7-point visual analogue scales ranging from "0%" to "100%" Mean (%) All interview participants believed that it is important for a preventive child healthcare professional to personally introduce himself to relevant partners, situated in the same area. Many said that these contacts have to be renewed regularly, and remembering the partners that they should have a low threshold of contacting the child healthcare, especially when suspecting child maltreatment. Although these percentages are far below optimal care (100%), they are a lot higher than expected, considering the extreme wording ("very easily"). It is likely that this is caused by self-overestimation and inability to estimate percentages. Therefore, these results should only be appreciated relatively to each other, instead of in an absolute sense. The most worrisome finding are the three professions at the top of the "That the parents will not attend the check-ups anymore" "That CPS will not help the family, after which they are returned to our care, but now with a damaged relationship" "Not being 'covered' by colleagues from child protective services / youth care" "I have not been confronted with this yet" (* experience 0 to 3 yr) "Not fear, but insufficient time/possibility to plan an extra check-up" "I never had to respond to maltreatment without cooperation of the parents'"