Facing suspected child abuse--what keeps Swedish general practitioners from reporting to child protective services?

OBJECTIVE: The aim of this study was to examine the reporting of suspected child abuse among Swedish general practitioners (GPs), and to investigate factors influencing them in their decision whether or not to report to child protective services (CPS).
DESIGN: A cross-sectional questionnaire study.
SETTING: Primary health care centres in western Sweden.
SUBJECTS: 177 GPs and GP trainees. MAIN OUTCOME MEASURES: Demographic and educational background, education on child abuse, attitudes to reporting and CPS, previous experience of reporting suspected child abuse, and need of support.
RESULTS: Despite mandatory reporting, 20% of all physicians had at some point suspected but not reported child abuse. Main reasons for non-reporting were uncertainty about the suspicion and use of alternative strategies; for instance, referral to other health care providers or follow-up of the family by the treating physician. Only 30% of all physicians trusted CPS's methods of investigating and acting in cases of suspected child abuse, and 44% of all physicians would have wanted access to expert consultation. There were no differences in the failure to report suspected child abuse that could be attributed to GP characteristics. However, GPs educated abroad reported less frequently to CPS than GPs educated in Sweden.
CONCLUSIONS: This study showed that GPs see a need for support from experts and that the communication and cooperation between GPs and CPS needs to be improved. The low frequency of reporting indicates a need for continued education of GPs and for updated guidelines including practical advice on how to manage child abuse.

The reporting of child abuse by Swedish general practitioners (GPs) at health care centres has so far not been investigated.

Swedish GPs did not report all cases of suspected child abuse to child protective services (CPS), despite mandatory reporting.

There is a need for better communication and cooperation between GPs and CPS, as well as continued education of GPs and updated guidelines including practical advice on how to manage child abuse.

Introduction

Child abuse, involving physical and/or psychological abuse, sexual assault, and/or neglect and failure to meet the child's basic needs is associated with major social problems and causes long-lasting consequences for the child [1,2]. Annually, 4–16% of all children in high-income countries are exposed to abuse [2]. Data from industrialized countries, including Eastern Europe, where the child abuse incidence is higher than in Sweden, show that around 10% of children seen by general practitioners (GP) have been exposed to abuse the preceding year [2,3].

In Sweden, physicians are obliged to notify child protective services (CPS) in cases of suspected child abuse [4]. Despite this, only 10% of notifications in 2012 originated from health care providers [5]. Several paediatricians and GPs at Child Health Centres (CHCs) in Sweden had never reported to CPS, and 67% failed to report suspected child abuse, according to a study performed in the late 1990s [6], which is in line with international results [1,7].

Reasons for not reporting suspected child abuse have been studied among paediatricians and GPs in the US and Australia [7-10]. Feelings, perceptions, and beliefs regarding child rearing and abuse affect the assessment [3,10]. Furthermore, psychological factors may prevent suspicion from arising [8,9]. Also, uncertainty about the “diagnosis” [1,7], lack of time [8], fear of offending parents and of destroying the relationship, thereby precluding continued monitoring of the child, may prevent the physician from reporting [1,7]. GPs, compared with paediatricians, were more afraid of losing contact with the family and waited for more evidence when they were uncertain about the suspicion, which reflects greater cautiousness with regard to reporting child abuse [10]. Previous experience of communication with CPS strongly influences the physicians’ reporting [1,7,10]. Physicians trained in detecting and reporting child abuse were more likely to suspect abuse than those who had received no training [11,12].

Physicians at CHCs in Sweden stated that reasons for not reporting suspected child abuse were fear of offending parents, uncertain observations, low confidence in CPS, and lack of time and training [6]. Also, in recent years an increasing number of physicians from other countries have been working in primary health care in Sweden and different culture and education might influence management of suspected child abuse [13]. Altogether, little is known about Swedish primary care physicians assessing and reporting child abuse and knowledge about their decision-making is therefore of importance.

The aim of this study was to examine the reporting of suspected child abuse among Swedish GPs, and to investigate factors influencing them in their decision whether or not to report to CPS.

Material and methods

Participants

The sample comprised all GPs and GP trainees (n = 177) in primary health care centres in Skaraborg, a rural part of Region Västra Götaland, Sweden. Unless otherwise indicated, both GPs and GP trainees are referred to as GPs in the following. The Regional Ethical Review Board in Gothenburg (Reg. no. 618-12) approved the study.

Variables

The definition of child abuse according to the Swedish Committee against Child Abuse was described in the questionnaire, including physical, sexual, and psychological abuse, and neglect/failure to meet the child's basic needs [14]. The questionnaire comprised questions on the GPs’ educational and professional background: educated Sweden/EU/other, education on child abuse (as a student/postgraduate), GP/GP trainee, private/public care, working years since education (< 10/11–20/> 20), working at CHC (currently/previously/never). Questions on availability of guidelines and perceived support at their clinic (yes/no/uncertain) were also included. Communication with CPS was evaluated (easy to contact or not, feedback or not). Experiences in the field were captured by number of cases of child abuse ever reported to CPS and failing to report despite suspicion (yes/no). Also, ever reporting to the police was questioned (yes/no). The statements relating to child abuse reporting had previously been used and evaluated [6,10]. A caution score ad modum Van Haeringen [10], novel for Swedish conditions, was constructed from four statements (see Table IV, questions 6–9). It was also possible to make free text comments. The questionnaire was piloted on six GPs with varying skills in the Swedish language.

Data collection

Data were collected anonymously using a web-based survey (esMakerNX2) in September/October 2012 [15], administered by a secretary not involved in the study. The computer programme automatically sent three reminders to non-responders.

Data analysis

Data were analysed using SPSS, the Statistical Package for the Social Sciences, version 20.0 [16]. Descriptive statistics described the sample and questionnaire data. A chi-squared test was used together with bivariate and multivariate logistic regression analyses to explore associations between independent variables (professional characteristics) and two dichotomized dependent variables: (i) ever reported (yes/no) and (ii) failing to report despite suspicion (yes/no). The caution score [8] (possible scores 1–5 from “strongly disagree” to “strongly agree”) was analysed in relation to professional characteristics.

Basic principles of content analysis were used to analyse the free text comments. In qualitative content analysis of a text, both manifest (visible and spoken statements) and latent (interpreted underlying meanings) content can be sought [17]. In the current study, manifest content is presented, including identification of meaning units (words/statements with related content), condensation (the meaning units are expressed in shorter phrases), abstracting/ formulation of codes (content areas of meaning units), and creating of categories (content with shared commonality).

Results

The questionnaire was completed by 77 GPs (44%). Table I shows the participants’ professional characteristics.

Table I.

Physicians’ professional characteristics.

	n	%
Physicians
GP1	49	64
GP trainee	28	36
Medical education
Sweden	44	57
EU2	27	35
Other	6	8
Working years since graduation
≤ 10 years	34	45
11–20 years	24	31
> 20 years	19	24
Working at CHC3
Currently	27	35
Previously	26	34
Never	24	31

1General practitioner.

2European Union.

3Child Health Centre.

Of the GPs who completed the questionnaire, 44 (57%) remembered having received education on child abuse during medical school. During the past five years, 36 GPs (47%) had received continued training on child abuse. Almost 25% of the GPs (18 persons) had guidelines on suspected child abuse at their workplaces, while 46 GPs were uncertain, and 13 reported lack of guidelines. Most GPs (95%) thought that the head of their clinic would support them in the assessment and reporting of suspected child abuse and 97% believed that discussions with colleagues would be possible.

The number of reports made by the GPs to CPS is shown in Table II. Forty-three (> 50%) of the GPs, including five GPs with more than 30 years’ work experience, had never reported suspected child abuse to CPS. Of the 37 GPs who at some point had reported suspected child abuse, only 17 GPs (46%) had received feedback from CPS. Nine GPs had been refused feedback from CPS. Only three GPs had ever reported to the police.

Table II.

Reports by GPs to child protective services during whole career and during the past year in relation to work experience.

Work experience		Number of cases reported during whole career				Number of cases reported during the past year
Years	n	0	1–2	3–5	6–10	0	1	2–3
0–2	3	3	0	0	0	3	0	0
3–5	12	7	4	1	0	9	3	0
6–10	19	14	3	1	1	18	1	0
11–20	24	11	9	4	0	19	4	1
21–30	8	3	2	3	0	8	0	0
> 30	11	5	3	3	0	11	0	0
Total	77	43	21	12	1	68	8	1

Fifteen GPs (20%) had failed to report suspected child abuse to CPS at some time. Reasons stated for not reporting are shown in Table III. The GPs’ attitudes towards reporting suspected child abuse to CPS are shown in Table IV. The mean scores for attitudes towards CPS were 3.1–3.2 (the neutral middle alternative). Most GPs disagreed with statements regarding a cautious approach to reporting suspected child abuse, resulting in a low caution score. There were no statistical differences in the caution score related to professional characteristics.

Table III.

Reasons for non-reporting of suspected child abuse stated by 15 GPs.1.

Stated reasons for non-reporting	No. of participants
Uncertainty about suspicion	6
Planned short-term follow-up of child	6
Referral to other health care provider	6
Child protective services already in contact with family	3
Lack of knowledge on child abuse	2
Fear of losing the family's trust and contact	2
Lack of time	2
Not expecting positive outcome for child when reporting	1
Could help the family on my own	1

Note: 1More than one reason could be indicated.

Table IV.

Respondents’ attitudes towards reporting suspected child abuse.

	Disagree n (%)	Neutral n (%)	Agree n (%)	Mean	SD
1. It's easy to contact child protective services	19 (25)	34 (42)	26 (34)	3.2	0.88
2. Child protective services do a good job in suspected child abuse	6 (8)	54 (70)	17 (22)	3.1	0.53
3. I trust child protective services’ investigations in suspected child abuse	13 (15)	41 (53)	23 (30)	3.1	0.86
4. I trust child protective services’ interventions in child abuse	14 (19)	39 (51)	24 (32)	3.1	0.80
5. Talking to families about abuse risks losing contact	14 (18)	32 (30)	31 (40)	3.3	0.97
6. I have a better chance of resolving maltreatment problems on my own	60 (78)	15 (20)	2 (3)	1.7	0.87
7. Reports should be made only if persistent pattern of abuse	80 (99)	0	1 (1)	1.1	0.49
8. Reports should be made only if quite certain of abuse	52 (68)	13 (17)	12 (16)	2.1	1.33
9. May be reasonable to defer reporting if no firm evidence, to maintain contact and learn more	47 (61)	15 (20)	15 (20)	2.3	1.18
Total “caution score”				1.8	0.66

Notes: Answers to the statements had options 1–5: “disagree” = option 1–2, “neutral” = 3, and “agree” = option 4–5. The mean values are based on the five-point scale. “Caution score” is the mean value of combined questions 6, 7, 8, and 9.

In the bivariate analyses, there were no statistically significant differences in failing to report suspected child abuse between GP trainees/GPs, long/short work experience, medical school in Sweden/abroad, guidelines or absence of guidelines, continued training or not, working in CHC or not, and a high/low caution score. Reporting of suspected child abuse was more frequent among GPs than among GP trainees (57% vs. 21%, p = 0.002), among those with work experience > 10 years than among those with shorter work experience (56% vs. 29%, p = 0.018), and among those educated in Sweden compared with those educated abroad (56% vs. 30%, p = 0.029). In multivariate logistic regression models, including all variables from the bivariate analyses, one single statistically significant independent variable for not reporting was identified; those educated abroad reported more rarely, OR 3.14 (95% CI 1.05–9.42). No significant variables were identified for failing to report despite suspected child abuse.

Facing suspected child abuse, 44% of the GPs indicated the need for support. In free text comments, the GPs defined the requested support as an accessible, competent resource for advice, for example a paediatrician, psychologist, or psychiatrist. Other comments concerned the GPs’ difficulties in discovering child abuse, communication problems with CPS, lack of information about CPS's way of working, and lack of feedback after reporting. Closer cooperation with CPS was requested. However, some GPs felt that the handling of suspected child abuse cases by the CPS was somewhat unprofessional. One GP expressed concern about his/her own or his/her family's safety and would like the possibility to report anonymously.

Discussion

In this study, one in five GPs at some time failed to report suspected child abuse. The most commonly reported reasons were uncertainty about the suspicion and the use of alternative strategies, such as referring the child to other health care providers or follow-up of the family by the treating physician. Less than one-third of the GPs trusted CPS's ways of handling suspected child abuse.

The response rate in the study was low, in line with similar studies [10]. The reason for not responding might be the sensitive topic of the study [18]. Possibly, those with a special commitment to the prevention of child abuse participated to a greater extent [18], while those who had failed to report such cases refrained from participating. The survey was conducted anonymously; thus, it was not possible to perform a follow-up of the non-responders and their characteristics are unknown. The low response rate resulted in small numbers in some subgroups; hence, true statistically significant differences may not be found (Type II error).

Self-reporting entails a risk of memory bias and social desirability [19]. Thus, the number of stated reports may differ from the actual number of reports. A qualitative approach; i.e. individual or focus-group interviews, might have provided more nuanced information.

A remarkably high percentage (60%) of the GPs did not know if guidelines regarding child abuse handling existed at their clinic. This may indicate little consideration of the possibility of child abuse and a lack of awareness of the need for guidelines. Future research might examine whether abuse is recognised through other members of the family and in non-CHC settings. The most common form of abuse, i.e. neglect, could be suspected in parents with for instance poor mental health and drug or alcohol misuse [20,21]. Fortunately, almost all GPs expected to receive support from their clinic heads and colleagues in suspected cases. Although many GPs experience stressful work situations [22], peer support might facilitate the discussion on and awareness of child abuse.

In our study, fewer GPs (44%) had reported to CPS compared with what was observed in studies of Australian GPs (72%) [10] and Swedish GPs working at CHCs in Gothenburg (77%) [6]. However, GPs working at CHCs meet more children than GPs in primary care, increasing the probability of meeting abused children. Also, they may receive continued education more often. The proportion of GPs who stated failure to report suspected child abuse in the current study (20%) was smaller than in the studies mentioned above, which were 47% [10] and 70% [6] respectively. An American study found rates as low as 8% [12]. Woodman et al. found low but increasing rates of recording of child maltreatment-related problems among English GPs [23,24]. Failure to report suspected child abuse is a sensitive subject and, in Sweden, a violation of the Social Services Act. Similar legislation exists in countries such as Australia and the US [10,25]. The GPs’ ability to recognize and report child abuse might differ between countries, depending on education, possibilities of multidisciplinary cooperation, and expert consultants. Also, the GPs’ cultural and religious background may influence the assessment of suspected child abuse [13]. However, all previous studies show that physicians do not report all cases of suspected child abuse [1,3,7].

None of the most experienced GPs in our study had reported suspected child abuse to CPS during the past year. Likewise, Australian physicians with longer work experience were less likely to report suspected child abuse than less experienced physicians [10]. Our data indicate that experienced GPs participated less frequently in continued education on child abuse. Studies have shown that the propensity to report suspected child abuse increases after training [8,12].

Uncertainty about the child abuse suspicion was a common cause for non-reporting, which is concordant with other studies [1,7]. Suspected child abuse implies difficult assessments, and studies show the importance of support, opportunities to seek advice from colleagues or experts, and having sufficient time for consultation and reporting [3,7,26]. Physicians who have access to expert consultation feel more secure in their decisions in cases of suspected child abuse [7,27].

Many GPs provided a neutral response to statements about CPS in the current study. Better communication and collaboration with CPS and possibilities of consultation may lower the threshold for reporting [28,29]. The total caution score for reporting in this study was lower than in the Australian study already mentioned [10], indicating a greater readiness to report among the Swedish GPs. However, the mean score for trust in CPS was the same in both studies. We need to further explore which factors cause the child abuse suspicion to arise and lead to reporting of the case [30].

Conclusions

This study showed that GPs see a need for support from experts and that the communication and cooperation between GPs and CPS needs to be improved.

The low frequency of reporting indicates a need for continued education of GPs and for updated guidelines including practical advice on how to manage child abuse.