Development and validation of a questionnaire to assess incidence and reactions of second victims in German speaking countries (SeViD)

Background Second victims, defined as health care team members being traumatized by an unanticipated clinical event or outcome, are supposed to be a common phenomenon in health care. Surveys in the US health care system indicate high incidence rates among physicians between 10 and 44%. However, no systematic assessment of second victims in health care in German speaking countries has been published yet and no validated German questionnaire for assessing incidence of and impact on second victims exists. Therefore, we initiated the SeViD (Second Victims im Deutschsprachigen Raum/second victims in German speaking countries) project and developed a German questionnaire for the assessment of second victim incidents. Methods Based on an intensive literature review of available questionnaires in English we defined a preliminary version of our questionnaire consisting of 4 domains and 14 items. This version was subject to cognitive pretesting using paraphrasing, probing and think aloud methods in order to ensure content validity. Retest reliability of second victim symptoms was assessed three weeks after the initial pretest. Results Fifteen health care professionals (physicians, nurses, therapeutic and diagnostic professions and administrative staff) of hospitals in Germany (n=6) and Austria (n=9) with or without previous second victim experience participated as volunteers for all pretests after informed consent. Seven items in three domains were slightly modified based on cognitive pretests. Retest reliability for second victim symptoms was rho = 0.76. Mean duration of completion for this questionnaire took 9:01 (±3:05) minutes in case of a previous second victim experience and otherwise 4:19 (±0:59) minutes and was regarded acceptable by all volunteers. No volunteer regarded any question to be inappropriate. Conclusion We successfully developed a validated questionnaire assessing the impact of the second victim phenomenon in inpatient health care facilities. This questionnaire will be used in different settings for health care professionals and for stand-alone baseline assessment as well as pre/post-survey along with complex educational interventions to reduce negative impacts of the second victim phenomenon.

literature review of available questionnaires in English we defined a preliminary version of our questionnaire consisting of 4 domains and 14 items. This version was subject to cognitive pretesting using paraphrasing, probing and think aloud methods in order to ensure content validity. Retest reliability of second victim symptoms was assessed three weeks after the initial pretest. Results Fifteen health care professionals (physicians, nurses, therapeutic and diagnostic professions and administrative staff) of hospitals in Germany (n=6) and Austria (n=9) with or without previous second victim experience participated as volunteers for all pretests after informed consent. Seven items in three domains were slightly modified based on cognitive pretests. Retest reliability for second victim symptoms was rho = 0.76. Mean duration of completion for this questionnaire took 9:01 (±3:05) minutes in case of a previous second victim experience and otherwise 4:19 (±0:59) minutes and was regarded acceptable by all volunteers. No volunteer regarded any question to be inappropriate.
Conclusion We successfully developed a validated questionnaire assessing the impact of the second victim phenomenon in inpatient health care facilities. This questionnaire will be used in different settings for health care professionals and for stand-alone baseline assessment as well as pre/postsurvey along with complex educational interventions to reduce negative impacts of the second victim phenomenon.

Background
Health care is associated with relevant risks not only for patients, but also for health care 3 professionals like infections or needle stick injuries [1][2][3][4][5][6][7]. Besides those well-known risks to physical integrity, unanticipated clinical events or outcomes, often caused by mistakes in health care, do not only traumatize patients but also health care professionals, who may become thus so-called second victims [8,9]. Being a second victim can lead to dysfunctional coping strategies [10] resulting in change in working behavior leading with reduced quality of care to further negative patient-and employee-related outcomes like isolation, reduced quality of life up to PTSD [10][11][12] or even suicide [13]. Previous surveys in English speaking countries indicate prevalence between 10 and 42 % of the second victims among health care professionals [14,15]. Based on research of the natural history of second victim traumatization [8] several interventional programs for health care professionals were launched in English speaking countries [16,20] showing beneficial evidence regarding employeerelated outcomes [16,17] and cost-effectiveness [18].
In Germany the association of statutory accident insurances defined standards for the care of employees after traumatizing events [19]. In opposite to sectors like rail services [21] or air traffic [22], where psychological support for employees after a traumatizing events has been addressed already, no systematic assessment of this phenomenon in the German speaking health care sector has been published yet.
We therefore initiated the SeViD (Second Victims im Deutschsprachigen Raum/second victims in German speaking countries) project. As a first step of this project we developed and validated a German-language questionnaire for assessment of second victim experiences as well as pre/postevaluation of intervention programs to reduce the impact of second victim incidents.

Methods
We conducted a systematic literature search in MEDLINE and Google scholar to identify questionnaires previously developed and/or used to evaluate the second victim phenomenon in health care, including only publications in English or German. We developed a German speaking version based on all identified questionnaires. In order to ensure comparability of surveys in different settings of health care we designed all questions in a multiprofessional way and did not intend to customize items to any medical specialty apart from demographic information. Duration of answering the 4 questionnaire was intended not to exceed five minutes in case of absence of a second victim experience to ensure acceptable response rates and to avoid selection bias. The preliminary version of this questionnaire was subject to cognitive pretesting using paraphrasing, probing and think aloud methods [23] in order to ensure content validity.
We included health care professionals of different professional groups (physicians, nurses, therapeutic  [23]. All volunteers received the same questionnaire three weeks after initial assessment to confirm retest reliability for the domain second victim symptoms [25]. Cognitive pretests were conducted by a researcher independent from both participating hospitals who was unknown to all participating volunteers to minimize the risk of observation bias.
To ensure standards of data protection, questionnaires were completely anonymized using respondent-generated codes to match tests and retests. Anonymized use of data for scientific reasons was declared in all questionnaires. Descriptive analyses were computed using Microsoft Excel © 2016.

Results
We identified six questionnaires related to nine resources to be potentially suitable for partial inclusion in our preliminary questionnaire [8,10,20,[26][27][28][29][30]31]. Details of included questionnaires are shown in Table 1. The first draft of our questionnaire consisted of the three domains general experience with second victim phenomenon, second victim symptoms and second victim support strategies and was limited to 40 items, that were taken over by or adapted from included questionnaires as shown in Table 2. For the symptoms domain participants answered by a 3-point (strongly pronounced, weakly pronounced, not pronounced) and for the support strategies domain by a 4-point (very helpful, rather helpful, rather not helpful and not helpful) ordinal scale. The options "Don't know" and "I cannot judge this", respectively, were also included.

5
Fifteen Health care professionals (physicians, nurses, therapeutic and diagnostic professions and administrative staff) of Charité, Berlin (n=6) and Hietzing Hospital, Vienna (n=9) participated in pretesting the preliminary questionnaire. Among all participating health care professionals 3 of 15 (20 %) had previous second victim experience(s) which is comparable with published prevalence from studies in English speaking countries [15].
Seven items in all three domains were slightly modified based on cognitive pretests. All participants were able to rephrase selected questions or to paraphrase technical terms like second victim correctly, even if they reported that they had never heard about the second victim phenomenon before our pretest. All participants completed retest of this questionnaire three weeks after initial assessment. Retest reliability for the domain second victim symptoms was acceptable with rho = 0.76. Mean duration of completion for this questionnaire took 9:01 (±3:05) minutes in case of a previous second victim experience and otherwise 4:19 (±0:59) minutes which was regarded acceptable by all volunteers. No volunteer regarded any question to be inappropriate or important information to be missing in this questionnaire.

Discussion
We were able to develop and pretest a questionnaire to assess second victim experiences in health care professionals in German speaking countries, that was regarded acceptable by all participating volunteers during cognitive pretesting. Although pretesting a small subgroup of participants is always associated with the risk of selection bias or Hawthorne effect, reactions of participating volunteers with previous second victim experiences indicate absence of mayor observation bias.
The selection of items were based on previously developed English questionnaires to ensure content validity also resulting in an acceptable level of retest reliability. Changes to the preliminary version were marginal and mostly related to effects caused by translation of existing questions to German.

Conclusion
We successfully developed a validated questionnaire assessing the impact of the second victim phenomenon in inpatient health care facilities. This questionnaire will be used in different settings for health care professionals for stand-alone baseline assessment as well as pre/post-surveys along with 6 complex educational interventions to reduce harm of the second victim phenomenon, like e.g. the

Funding
This research received no specific grant from any funding agency in the public, commercial, or notfor-profit sectors.

Availability of data and materials
The questionnaire is set up in the German language and therefore not suitable for international use.
Interested researchers are encouraged to contact the authors for the provision of the used questionnaire.

Authors' contributions
RS, MA, WH, BE and MR conceived the study. RS, MA and MR collected data and provided the first draft. All authors read and approved the manuscript.

Ethics approval and consent to participate
Because of the research design, no formal vote of the Ethics Committee was required. To ensure data protection all data were collected without any demographic information allowing identification of participants. All participants gave their consent to use of data for this study.

Consent for publication
All participants were informed about the study and gave their consent to publication of survey data.