Physicians’ Attitudes toward Adolescent Confidentiality Services: Scale Development and Validation

A cross-sectional survey was conducted in the period from November 2017 to August 2018. Data were collected from a sample of physicians working at school pediatric departments and gynecology departments (from the age of 15 adolescent girls are referred to the gynecology department at the primary care level) in 13 primary healthcare institutions in the territory of Belgrade, Serbia. There are 16 primary healthcare institutions in Belgrade, but 3 institutions declined to participate in the study. We chose to conduct our study at the primary healthcare level because it has the main role in health prevention and promotion. Primary care physicians are in a position to establish relationships of trust with adolescent patients, and to influence them to seek advice and care for sensitive health issues such as sexuality and reproductive care. The study protocol was approved by the Ethics Committee of the Faculty of Medicine, University of Belgrade (approval number: 29/ VI-1).

Participants were approached at their work offices by the principal researcher after receiving approval from the heads of the departments. Questionnaires were self-administered and anonymous. The survey took approximately 10 minutes to complete. All participants signed informed consent sheets prior to the survey.

The items developed for the scale were based on a qualitative interview study exploring primary care physicians’ knowledge, beliefs, attitudes and experiences regarding their legal obligation to protect adolescents’ confidentiality in healthcare (20). The qualitative study included 12 interviews with primary care pediatricians and gynecologists. Some of the categories obtained by inductive qualitative content analysis were the basis for the key constructs in the study of confidentiality attitudes and main domains in this survey’s questionnaire. Categories that served as a basis for questionnaire development were: respect for the person, condition for trust, clinical situations where confidentiality is most important, parent-adolescent sexuality communication, decision-making capacity of adolescents, and parental responsibility. In accordance with the relevant literature on adolescent confidentiality, it was expected that appreciation of the main reasons for confidentiality (respect for autonomy and establishing doctor-patient trust), a positive attitude toward confidentiality in specific clinical situations, positive views on the maturity of adolescent judgment, and recognition of the adolescent’s need to keep their reproductive and sexual health problems private from parents, would contribute to an overall positive attitude toward respecting adolescents’ right to confidentiality in reproductive healthcare (21, 22). Items of the questionnaire were partially formulated employing phraseology from physicians interviewed to generate a richer understanding of target respondents.

Content validity of the instrument was established in the following steps. First, the main concepts were defined based on a literature review and the results of the qualitative study. Second, a panel of 5 experts in a content domain (one expert from each of the fields of psychology and public health) were sent a questionnaire and were invited to evaluate the items. Experts were presented with constructs and asked to match items with a corresponding construct. They confirmed that the instrument items were relevant, accurate, and adequately represented the theoretical constructs that the questionnaire was designed to measure. Third, the questionnaire was piloted in a group of 10 primary healthcare physicians to assess the intelligibility of the questions. The instrument was suitable for all participants, so no adjustments were necessary.

The final version of The Attitudes toward Adolescent Confidentiality Scale consisted of 20 statements (items of the scale) which represent various constructs related to the right to confidentiality in adolescent healthcare, evaluated on a 5-point agreement scale (from 1 “Strongly disagree” to 5 “Strongly agree”) (Table 1). Reverse coding was applied for negatively connoted statements (A8, A9, A10 and A20 in the Table 1.). The attitude score was calculated by summation of points for each item.

Principal component analysis with varimax rotation was applied to determine the main components (factors) representing constructs that the scale is supposed to investigate, and items with factor loadings less than 0.3 were removed. Because no validated scales of similar constructs were found, construct validity was assessed by calculating the correlation coefficients between each pair of subscale scores. The Kaiser-Meer-Olkin measure of sampling adequacy and Bartlett test of sphericity were calculated.

The reliability of the scale was assessed by calculating the Cronbach alpha coefficients and mean inter-item correlations. Cronbach alpha was calculated for each factor, too. Test-retest reliability analysis was impractical because Serbian primary healthcare physicians were reluctant to complete the survey again due to the ongoing significant work overload in Serbia. Furthermore, a split-half reliability test was performed and Spearman-Brown coefficient calculated.

Descriptive statistics were applied to determine the sample characteristics. Mean scores for the total scale and sub-scales were calculated. Mean scores for individual items of the scale were calculated to assess the importance our respondents attach to particular constructs related to adolescent confidentiality (23). In all analyses p < 0.05 was considered statistically significant. All analyses were performed using the statistical program IBM SPSS Statistics 20 (SPSS).

The first factor included items related to the attitude of respect for confidentiality in specific clinical situations. Its reliability was very good (Table 3). The mean score on this subscale was moderate, with higher scores for items related to preventive health measures (counseling and contraception) (Table 3). Lower scores were observed on items related to abortion and pregnancy.

The second factor covered items related to the perception of the importance of confidentiality in adolescent healthcare (ethical duty, respect for persons, condition for trust and condition for healthcare service access). This subscale also had very good reliability (Table 3).

Respondents’ scores on this subscale were the highest scores, indicating they had a positive attitude towards the reasons that support confidentiality policies (Table 3). The third subscale included items related to the maturity of adolescents and parental responsibility. This subscale had satisfying reliability (Table 3). A negative average attitude was found on this subscale (Table 3).

The fourth factor covered items related to the issues of communication and secrecy between adolescents and parents regarding sexuality and reproductive health. Communication with parents subscale had a Cronbach alpha below the acceptable level (Table 3), but its mean inter-item correlation was sufficient to keep this subscale in the instrument. Further work with developing the concept of secrecy in the adolescent-parent relationship and adding more items to this subscale are needed to improve this domain of the scale, and only the total score should be used for the time being. Respondents’ scores on this subscale were moderate (Table 3).

In this article we describe the development and validation of a research instrument aimed at measuring physicians’ attitudes towards adolescent confidentiality in healthcare. This study used a systematic approach to develop the instrument, and thoroughly analyzed the instrument’s reliability and validity. The overall reliability of the 19-item Attitudes towards Adolescent Confidentiality Scale was acceptable, showing high consistency of the total scale (Cronbach alpha of 0.83). Principal component analysis showed four components, which present subscales of the instrument: Confidentiality in a specific situation, the Importance of confidentiality, Adolescent maturity, and Communication with parents (Table 3).

The developed scale also showed a satisfying level of construct validity, given that the subscales correlated in expected ways (Table 2). Physicians’ higher acceptance of confidentiality in specific medical situations was associated with stronger beliefs that confidentiality is important, as well as with more positive assessments of adolescent maturity and decision-making capacity. Also, higher acceptance of the importance of confidentiality was associated with stronger belief that adolescents keep secrets from parents and do not want parents to find out. As expected, the attitude toward adolescent maturity and parental responsibility was not significantly correlated to recognition of adolescents’ need to keep their sexual and reproductive health secret from parents, since those subscales present significantly different constructs. A negative attitude toward the maturity of adolescents’ judgment and a positive attitude toward parental authority in healthcare decision-making underlie the strongest arguments against adolescent confidentiality policies in the literature (5, 6). Appreciation of the fact that adolescents would rather forgo needed sensitive healthcare because of the fear that parents might find out led to the formulation of confidentiality policies which aim to increase adolescents’ access to healthcare services (27,28,29).

The surveyed physicians demonstrated a moderate overall attitude toward the adolescent’s right to confidentiality in healthcare, which was expected in Serbian physicians, given that the general attitude toward children’s rights in the general Serbian population is predominantly negative (30). The moderate attitude reflects physicians’ doubts and confusion and suggest that confidentiality protection in adolescent healthcare may not be consistently implemented. This result is in line with the empirical evidence showing that the socio-cultural context strongly influences the implementation of minors’ rights (31). In societies with dominant patriarchal cultures, physicians are less inclined to promote children’s autonomy in healthcare (32). An inter-cultural comparison of results would be important in future explorations of socio-cultural influences on the implementation of minors’ participation rights in healthcare. Empirical evidence from the U.S. shows that despite recommendations and legal requirements, less than half of adolescents aged 15–17 spent some time alone with a physician during a visit (33, 34), implying that the importance of adolescent confidentiality is not universally accepted.

As expected, scores on the subscale Confidentiality in specific situations were higher for items related to preventive health measures (counseling and contraception) (Table 3). Lower scores were observed on items related to abortion and pregnancy, which was expected and consistent with the results of other studies that suggest that a significant number of physicians believe that adolescent pregnancy and abortion should never be kept confidential and parents should be included in making decisions in these situations (18, 35). Respondents’ scores on the Importance of confidentiality were the highest scores (Table 3), which is in line with the results of other studies where a majority of physicians agreed with the importance of adolescents’ right to confidentiality (18, 35). Despite recognizing its importance, the physicians surveyed inconsistently maintained confidentiality in their practice (18, 35). A negative average attitude was found on the Adolescent maturity subscale, which implies that many physicians disagree with legal regulations that allow 15-year-old patients to independently consent to medical procedures (Table 3). Similar findings were found in a survey by Riley et al., where respondents, although they agreed that confidentiality should be an adolescent’s right, were less confident that adolescents possess adequate maturity to independently consent to care for sensitive medical issues (18). Healthcare practitioners often have a patronizing view of adolescent patients, believing they are immature and in need of protection (36). These findings reflect the main dilemma surrounding the issue of confidentiality in adolescent healthcare: Are adolescents mature enough to make sound autonomous decision in healthcare? We argue that this question should be put aside in the field of reproductive health, and the emphasis be put on public health justification that is based on the health benefits of adolescents getting needed medical help. Physicians’ negative attitude toward adolescent maturity jeopardizes adolescents’ access to healthcare, especially in areas of reproductive and mental health. Thus it is important to address physicians’ beliefs and prejudices in order to tailor educational interventions.

Respondents’ scores on the Communication with parents subscale were moderate, which indicates a lack of consensus on this issue among the population of surveyed physicians (Table 3). However, the results of numerous studies with adolescents showed that many adolescents value confidentiality highly in healthcare and prefer that parents not be informed about their reproductive health issues (contraception, sexually transmitted diseases) (7, 33, 37, 38). Insisting on parental notification would only discourage adolescents from seeking necessary medical care.

There are important limitations to this study. First of all, our sample of primary care physicians was small, although sufficient for running principal component analysis (39). For the purpose of exploratory research, we decided to include only physicians working in Belgrade, and their number was not large. Still, data collection took several months. Excess workload and continued health professional emigration that is going on in Serbia were the main obstacles to instrument administration, and this also precluded test-retest analysis. The scale should be validated on a larger sample of physicians, since skewness and kurtosis suggest that distribution in our sample was not normal (Table 4). Physicians from both urban and rural areas should be included to examine the influence of geographic and socio-cultural factors on physicians’ acceptance of minors’ rights in healthcare. Furthermore, it would be useful to validate translated versions of the instrument in different countries and cultures. We invite the broader professional community to adapt and validate the scale for their populations. Secondly, the majority of participants in our sample were females (83.6%). According to the report of The Statistical Office of the Republic of Serbia, gender ratio in the total sample of Serbian medical doctors is 2:1 (for female and male doctors, respectively) (40). According to the World Health Organization’s evaluation of primary healthcare in Serbia, primary care is dominantly provided by women (41). In this evaluation 89% of respondents were women. So the sample in our study adequately represents the gender ratio of primary healthcare in Serbia.

Thirdly, we did not include private healthcare institutions. Future studies are needed to explore differences in attitudes and approaches in adolescent healthcare between public and private clinics. Fourthly, our sample was sufficient for the principal component analysis as an initial exploration of the instrument’s factor structure. Confirmatory factor analysis in a different sample is needed to evaluate the factor structure.