Research

I completed my research dissertation at the Medical University of South Carolina August 1, 2018. The dissertation is a compendium comprised of 3 published/publishable manuscripts. These three manuscripts examine different aspects of adolescent risk screening in the primary care setting. One has been published and the other 2 are ready for submission. Each manuscript offers a different view of the challenges associated with this facet of adolescent care. All the studies reflected in the three manuscripts utilized the Donabedian Framework of structure, process, outcome to evaluate the designated components of adolescent risk screens and adolescent risk screening. A summary of each manuscript follows.

Manuscript 1

The first manuscript evaluated the literature to review adolescent risk screens available for use in the primary care setting (Hiott, Phillips, & Amella, 2017). There are few multidimensional or multi-problem risk screening instruments available. Multidimensional screens assess for several risk behaviors; this is important as research demonstrates that most adolescent risk behaviors co-occur (Aspy et al., 2012).

When the structure of the multidimensional risk screens was examined using the Donabedian Framework, several problems were identified. These problems included: cost, the number of domains assessed, length of the screen, format of the questions (‘yes’/’no’ or Likert), and ease of scoring.

Another concern involving adolescent risk screening in the primary care setting involved the process of assessing for risks. Some screens used parent report rather than self-report. Studies indicate that internalizing behaviors such as depression are often missed by parents and that adolescents are less likely to admit to risky behavior in the presence of parents or guardians (Herrera, Benjet, Méndez, Casanova, & Medina-Mora, 2017; Pagano, Cassidy, Little, Murphy, & Jellinek, 2000).

Outcomes of the evidence examined the strength of the studies that examined the instruments and the validity and reliability of the screens examined. Psychometric properties were unavailable for several studies. Consequently, when recommending adolescent risk screening surveys, the Pediatric Symptom Checklist (PSC) and its derivatives are validated, reliable and brief. This instrument is free in the public domain and available in the Bright Futures Mental Health Toolkit.

Manuscript 2

Manuscript 2 examined a common risk screening interview used in pediatric and primary care settings called the HEEADSSS. HEEADSSS stands for home, education, eating, activities, drugs, sexual activity, suicide (depression), and safety. The HEEADSSS interview is used worldwide for identifying adolescent risk behaviors has been called the ‘gold standard’ of psychosocial interviews, and yet little is known about its psychometric properties (Sturrock & Steinbeck, 2013). The Donabedian Framework of structure, process, outcome was used to examine what is known about the HEEADSSS interview.

The structure of the HEEADSSS was examined by evaluating the various iterations of the HEEADSSS that have been studied. The acronym has been added to over time with the addition of new letters and new domains and some versions of the HEEADSSS had other words substituted for some of the letters. This review of the literature examined settings that used the HEAADSSS, documentation rates, completion time, as well as the purposes of the studies.

The process of using the HEEADSSS was evaluated in only 3 studies; these evaluated using the HEEADSSS in a written or computerized format. No studies evaluated the provider or adolescent perceptions, comfort levels or satisfaction with the face-to-face, self-report format.

In this literature review outcomes were assessed by examining any reliability or validity scores reported and documentation of completion rates. No reliability or validity scores were discovered for the HEEADSSS interview and completion and documentation rates of the results of the HEEADSSS interview were low. Studies revealed that documentation of the HEEADSSS interview averaged about 10% with 20% to 28% of adolescents not being asked any HEEADSSS questions at all.

This review of the evidence concerning the HEEADSSS interview was compiled to afford health care providers current, relevant information about this often-used instrument. It is easy to remember but can take 5 – 40 minutes to complete. Use of this interview has not been evaluated psychometrically. Additionally, no studies have examined the perceptions, comfort levels or satisfaction of providers or adolescents concerning this often-used self-report, face-to-face interview.

The evidence surrounding multidimensional risk screens available for use in the primary care setting and the HEEADSSS interview demonstrate gaps in preventive care for the adolescent population. In order to explore process improvements regarding adolescent risk screening within the primary care setting a feasibility study was devised to examine adolescent and provider interest and acceptability of using the HEEADSSS interview combined with the Pediatric Symptom Checklist – Youth (PSC-Y) and the Issues Checklist (IC).

Manuscript 3

Manuscript 3 reviews the feasibility study that was implemented. The objective of this feasibility study was to examine adolescent and provider interest and acceptability of using the Pediatric Symptom Checklist – Youth (PSC-Y) and the Issues Checklist (IC) administered via a computer tablet as instruments to complement the HEEADSSS interview. Agreement was assessed between all the instruments used with Bland-Altman plots and the Donabedian Framework was used to evaluate the structure, process and outcome of the feasibility study.

Adolescents felt that the study was important and wanted to help. They felt the surveys were easy and relevant. Providers were surprised that adolescents were so receptive and pleased that it did not interfere with their schedules. Another structural element was the use of the tablet, adolescents preferred using a tablet or phone over written surveys. Yet, when it came to the risk screening process, while they enjoyed talking with their provider, they felt less comfortable speaking with them about personal matters.

Providers agreed that improvements need to be made in adolescent risk screening in the primary care center. They agreed that it would be very helpful to the risk screening process if the adolescents could complete the surveys before their visit and felt that the use of technology could be very useful.

Providers were less enthusiastic about the screens used, the Pediatric Symptom Checklist – Youth (PSC-Y), the Issues Checklist (IC) and the HEEADSSS. Generally, the providers felt the IC was long. Some providers liked the PSC-Y but felt that they did not get as clear a picture of the specific psychosocial risk areas as they did with screens that assess for only one risk domain. However, other providers felt that the PSC-Y would be helpful to determine the adolescents state of mind.

Reviews for the HEEADSSS were mixed as well. Providers acknowledged that the HEEADSSS did prompt them to ask adolescents about some topics they do not routinely address, and the interview did produce the needed outcome of risk identification. Providers voiced concerns about structure of the HEEADSSS interview. They noted that it is so loosely structured that topics could be asked in a closed format eliciting a ‘yes’ or ‘no’ answer, topics could be ignored, or time could run out before the interview was finished. This would ultimately interfere with the risk screening process.

Statistically, when the PSC-Y, IC, and the HEEADSSS interview were evaluated using the Bland-Altman plot, agreement was found. This indicates that similar numbers of risk were uncovered and that the PSC-Y and the IC could potentially substitute for the HEEADSSS or vice-versa. Since the areas of risk reviewed by each screen differ, this could reveal other concerning behaviors that need to be discussed during the HEEADSSS interview. Consequently, a logical choice would be to use these instruments to complement the HEEADSSS interview by administering them before the visit. Interestingly, the PSC-Y and IC intensity score were correlated. More research would need to be completed to determine what this means.

Significance and Future Implications

The culmination of these studies did reveal the need for innovation with adolescent risk screening in the primary care setting. Free, short, easy to score, valid and reliable risk screens that are self-report and offer Likert responses are needed in a format that adolescents feel is easy and confidential. Valuable information regarding the HEEADSSS interview revealed this interview is a practical instrument that can be easily incorporated into the adolescent visit. However, its usefulness may depend upon the provider’s ability to use the interview as it was designed, as well as the adolescents’ level of trust. As suggested when it was developed, the HEEADSSS should be used in addition to other risk screening measures (Goldenring & Rosen, 2004).

The PSC-Y and the IC are both validated and reliable screens that can indicate the need to further question adolescents about risk behaviors. The Donabedian Framework of structure, process and outcome provided an organized and logical way to evaluate screening methods, screening instruments, and adolescent and provider interest and acceptability especially in the primary care setting. Future implications include continued examination of appropriate risk screening instruments delivered in innovative ways to engage adolescents.

References

Hiott, D. B., Phillips, S., & Amella, E. (2017). Adolescent risk screening instruments for primary care: An integrative review utilizing the Donabedian Framework. Comprehensive Child and Adolescent Nursing, 21. doi:10.1080/24694193.2017.1330372