Quality Improvement: A Template-Based Approach to HEADSS Assessments for Hospitalized Adolescents Free

This study took place from September 2020 to December 2021 at 2 sites within a single university health system. The sites have a combined total of 74 pediatric floor beds with approximately 2500 annual medical admissions to pediatric floors, 25% of which are 13 years or older. For the adolescent age range considered, greater than 90% of patients have at least 1 chronic medical condition and many have complex medical needs. One site includes an adolescent eating disorder medical stabilization program. Care teams at both sites include the same pool of pediatric residents, medical students, fellows, and attending physicians. Most admissions to medical teams on a pediatric floor are performed by a pediatric resident or medical student with appropriate supervision, whereas a small portion are performed by a pediatric hospital medicine attending alone.

Patients aged 13 years and older admitted to a pediatric floor under a medical team were the focus of the interventions and were included in the analysis. There is no age at which a HEADSS history is a universally appropriate tool, and this age range was chosen because, in our experience, most of these patients can meaningfully engage with the history-taking process and some will have positive findings. Patients were included irrespective of preferred language. Patients admitted to surgical teams were excluded because interventions would be logistically difficult (providers on surgical teams use their own note templates and may have no experience with HEADSS histories) and because less than 5% of patients at our institution are admitted to surgical services. Patients admitted to an ICU were also excluded because many would be too ill to participate. Patients with conditions precluding their ability to provide a history were not excluded because we could not always accurately identify these patients during chart abstraction and attempts to do so could introduce bias.

Our quality improvement working group included representatives from Pediatric Hospital Medicine, Adolescent Medicine, and the Pediatric Residency Program. Based on clinical experience and discussions with key stakeholders, the group identified key drivers of incomplete HEADSS histories (Fig 1), which included providers forgetting to discuss and document a HEADSS history, providers not considering the HEADSS history to be a routine part of their inpatient practice, time constraints, lack of awareness regarding confidential notes, and situations (particularly COVID-19) that prevented speaking to patients privately. Informed by 6 months of baseline data (October 2020–March 2021), the working group set a SMART aim of improving the rate of complete HEADSS histories during admission for patients aged 13 years and older admitted to a pediatric floor from 11% to 31% within 8 months of the initial PDSA cycle. A complete HEADSS history was defined as documentation of information from all 6 domains. Histories documented under Social History or History of Present Illness were counted if they included information from a HEADSS domain.

At the start of April 2021 (immediately following the 6-month baseline period), the standard pediatric admission note template was modified to include a new prompt consisting of an instructive component directed at providers and a note component to facilitate documentation. The instructive text reminded providers to, for patients aged 13 years and older, document a HEADSS history in a confidential note separate from the main admission note using a template created by the working group. These instructions were formatted so they could be easily deleted and would not accidentally become part of the final note. The confidential note format existed before our study and use of it was considered a best practice, but discussions with stakeholders revealed that few providers were using or aware of this option. The note component of the prompt allowed providers to choose among several options for documentation: a statement that a HEADSS history was documented in a confidential note, a statement that the patient was younger than age 13 years along with a template for a focused social history for that age, or a statement that a HEADSS history could not be obtained at that time with a drop-down list to select a specific reason why along with a template for a focused social history.

In April 2021, we implemented a note template developed by the working group to facilitate documentation of a HEADSS history within the confidential note. This template simplified documentation and reminded providers of the 6 domains included in a complete HEADSS history. The template used cascading logic and, if certain findings were positive, provided clinical decision support for offering interventions and resources.

Educational efforts were made at the time note templates were implemented in April 2021. Face-to-face education was done with providers working on pediatric floor teams. Instructive fliers with contact information for working group members were placed in provider workrooms. Educational information was e-mailed to all pediatric residents and pediatric hospitalists. A working group member gave a brief presentation to pediatric residents during a noon conference. Initiatives included information about the goals of the project, changes being made, and benefits for patient confidentiality.

In May 2021, 1 month after the implementation of note template changes, working group members elicited open-ended feedback from providers who had worked in the inpatient setting during the previous month. This feedback led to changes to streamline the note templates including advisory text on how to refer high-risk patients to an adolescent medicine clinic.

In August 2021, given the new academic year, an educational e-mail was sent to all new pediatric interns informing them of the ongoing quality improvement initiative. An e-mail was sent to all pediatric residents sharing preliminary results of the initiative to remind residents of the project and provide positive reinforcement for the changes being made.

Six months of baseline data (October 2020–March 2021) were collected. Following the initial intervention at the start of April 2021, data were collected on a monthly basis. A data collection tool was developed and reviewed by all authors. Data were collected by chart review by a single author and stored in a secure electronic format. All patients who were eligible for the study were included for analysis.

The primary outcome measure was the percent of patients aged 13 years and older admitted to a medical service on a pediatric ward who had a complete HEADSS history (including information from all 6 HEADSS domains) documented at admission. Process measures included the percent of patients each month who had a confidential note used in documentation, percent of patients each month with sexual history documentation, and average number of HEADSS domains documented. Documentation of sexual history was included as an indicator of how frequently providers were engaging with more sensitive topics because, in our discussions with providers, they were often concerned about the sexual history when it came to confidentiality and documentation. There was consideration that the prompts in the note template would simply be deleted by providers, leading to notes without any social history documentation, so the percent of admission notes without social history documentation was a balancing measure. Study measures were analyzed using run charts. A shift was defined as 6 or more consecutive points above or below the center line, whereas a trend was defined as 5 or more consecutive points either all increasing or all decreasing.²⁵  Demographic data were collected, and χ² tests, paired t tests, and Fisher exact test were used as appropriate to compare the pre- and postintervention groups. Race and ethnicity were considered because these social constructs could influence history taking or documentation. Admission for an eating disorder was considered because, in our experience, these patients often have more complete social history documentation. Microsoft Excel with QIMacros was used to create run charts and to perform statistical analysis. Residents were surveyed to collect feedback and ideas for improvement, which informed subsequent interventions.

This project was deemed to be quality improvement by our institutional review board and thus did not require formal approval in accordance with our institution’s policies.

A total of 539 admissions were included, consisting of 212 in the baseline period (October 2020–March 2021) and 327 in the intervention period (April 2021–November 2021). There were no significant differences between the baseline and intervention population characteristics (Table 1).

The run chart for the primary outcome measure (Fig 2) shows the rate of complete HEADSS histories increased from 11% to 39% of admissions, with an upward shift during the improvement period. Run charts for process measures (Fig 3) show the use of a confidential note increased from 14% to 38%, documentation of a sexual history increased from 18% to 44%, and average number of HEADSS domains documented increased from 2.2 to 3.3, with an upward shift during the improvement period for each. In cases in which a confidential note was not used, the history was either documented in the admission note or not documented. The run chart for the primary balancing measure (Fig 4) shows no shift or trend in the percent of patients with no social history documented.

Through changes to an admission note template, creation of a HEADSS history template with clinical decision support, refinement of these templates based on provider feedback, and educational initiatives to support implementation, we were able to increase the rate of complete HEADSS assessments. PDSA cycle 1 led to a significant and sustained increase in the primary outcome measure that surpassed the SMART goal. Subsequent PDSA cycles 2 and 3 did not lead to further changes in the outcome measure but may have contributed to the sustained change observed. The success of PDSA cycle 1 highlights how creating structural changes to a workflow that meet the specific needs of local providers can result in sustainable change. Discussions with stakeholders revealed that many providers saw value in taking and documenting a HEADSS history but either forgot to do so or had concerns regarding confidentiality. This informed the design of the templates that were further refined based on provider feedback, which increased the likelihood that desired practices would be adopted. There was no decrease in outcome or process measures in July 2021, though new interns had started and had not received any education from the working group, suggesting the templates had become an accepted part of the workflow. Also, the smaller refinements to the templates in PDSA cycle 2 did not produce change even though they were based on specific feedback from providers. This suggests not all changes to workflow will necessarily be additive, and smaller, targeted changes may be less effective than larger, more fundamental shifts. Educational efforts are necessary in the implementation of many quality improvement efforts but rarely result in sustainable change in isolation. PDSA cycle 3 was primarily educational and did not result in change, though it may have contributed to sustainability.

There was also an increase in the use of confidential notes. It is possible this contributed to increased HEADSS documentation if providers felt more confident in the safety of documenting sensitive topics because this concern was brought up by many providers. This is supported by the increase in sexual history documentation, which suggests providers were increasingly discussing the more sensitive HEADSS topics. This is especially important in a health system caring for many patients with chronic illnesses because providers may incorrectly assume that chronically ill children are not sexually active. Alternatively, it may simply reflect provider engagement with the admission note template prompts. In either case, use of confidential notes enhances protection of sensitive patient information.

It is possible that an increase in complete HEADSS histories could occur even if a significant subset of patients had less overall social history documentation, which would be counterproductive. However, the increase in the average number of HEADSS domains documented suggests the change in the primary outcome measure was truly driven by an increase in the breadth of social history documentation.

The primary balancing measure, percent of admission notes with no social history, was unchanged from baseline. This project, like many quality improvement projects, asked providers to do more with the limited time they have, and its success depended on many different providers adapting their practices. That providers did not simply delete the prompt and omit the social history altogether suggests they saw such documentation as valuable, and the templates fit into their workflow without disruption. Any intervention that cannot secure buy-in from key stakeholders or that is too disruptive to existing practices is unlikely to result in sustainable change.

Although this study was designed based on local factors, it is plausible that many of the drivers (especially those related to provider memory and concerns about patient privacy) are present at other institutions. Others may find this experience instructive when attempting to achieve similar changes. For example, our results are consistent with another inpatient study using note template prompts to promote confidential note usage, suggesting prompts can be effective if provider memory is a key driver.²⁶  Another previous education-based study led to improvement in social history documentation but with diminished effect over time, mirroring the lesser impact of our PDSA cycle 3.²⁷  Another study that increased sexual history documentation found note template changes can replace repeated educational reinforcements when sustaining improvement.²⁸  Our results build on these studies by demonstrating sustained, improved documentation across a wide range of social domains. We note that our interventions were implemented without direct, upfront financial cost to the health system, so there is a low barrier to entry for others to implement a similar process. It is possible costs were incurred because of increased utilization of services such as social worker evaluations or testing for sexually transmitted infections, though, although this study was not designed to assess this, those costs may be reasonable if they create value for patients or prevent the need for more expensive future interventions. Likewise, changes that become a routine and expected part of the workflow and institutional culture may be more easily self-sustaining without costly effort from a health system.

This study has some limitations. As a single health system quality improvement initiative, it may not be generalizable. Our data include patients who were not cognitively able to participate in a HEADSS history because chart review often cannot determine precisely why a HEADSS history was not performed. To limit selection bias, we included all patients in the target age range, so the true rate of complete HEADSS histories is higher than reported here. A next step would be to directly evaluate a cohort of patients to determine a more accurate denominator of patients who could feasibly have a HEADSS history recorded. Furthermore, this study took place during the COVID-19 pandemic, and the frequency of adult caretakers not being able to leave the hospital room because of infection control measures, which would preclude a confidential HEADSS history, may have affected the baseline and intervention periods asymmetrically.

Improved note templates led to a sustained increase in complete HEADSS histories and use of confidential notes. Additional revisions to templates, repeat education, and providing project data to residents led to sustained improvement without additive impact on outcome measures. Higher rates of inpatient screening are expected to lead to more interventions for patients.²⁰  Others seeking to increase inpatient HEADSS histories may find this experience valuable. Future studies may delineate why some patients do not have a HEADSS history completed, determine if clinical decision support leads to appropriate interventions, and address barriers to screening patients on surgical teams, admitted to intensive care units, and who were not screened on admission to the floor.

The authors acknowledge the faculty, residents, and medical students who supported this project through their clinical work, feedback, and willingness to change practices.