HINF 510 Week 3 Key Design Elements for a Clinical Information System Example

Shorter Notes, Fewer but Better Alerts and One Medication List: A Design Matrix for a Health Center Network's New Electronic Record

[Student Name]

University of Phoenix

HINF/510: The Systems Life Cycle

Week 3 Assignment

[Instructor Name]

[Date]

The health center network, its design choices and data are composites written for a model paper; research findings come from the sources listed.

When the design sessions for the health centers' new record began, a family physician asked for fewer pop-up alerts and the quality director asked for more reminders to raise cancer screening rates. Both were right. The session made clear that design decisions, not the software itself, would determine whether the new record helped or hindered care. This paper sets out eight key design elements, the options considered for each and the decisions made, in the form of a design matrix.

How Decisions Were Made

Each element was decided by a design group of clinicians, analysts and the vendor's consultant, using the nine requirements from the systems assessment, research evidence and testing with users. The three clinical chiefs, for medicine, dentistry and behavioral health, signed off on the final matrix.

Element One: The Shared Clinical Record

Options: separate modules for medical, dental and behavioral health linked by interfaces, or one record with shared lists. Decision: one record with shared medication, allergy and problem lists visible to all three disciplines, with discipline-specific note types. Rationale: the assessment found split records caused duplicate prescribing, and integrated care is the network's model.

Element Two: Documentation Templates

Options: comprehensive templates that capture every possible finding, or short problem-based templates with free text. Decision: short templates organized by the patient's problems, with limited auto-inserted data and no copy-forward of whole notes. Rationale: in a study of 2.7 million outpatient notes at one academic center, median note length grew 60.1%, from 401 words in 2009 to 642 in 2018, redundancy rose, and in the final year fewer than a third of the words in a typical note had been typed fresh by its author, with the remainder templated or copied (Rule et al., 2021). Longer notes take longer to write and to read, and redundant text hides what matters.

Element Three: Clinical Decision Support

Options: activate the vendor's full library of alerts, or build a small set of targeted interventions. Decision: a small, targeted set. Rationale: a systematic review of 70 trials found decision support improved practice in 68% of them, and 30 of 32 systems with four features improved practice: support provided automatically within the workflow, recommendations rather than just assessments, support delivered at the moment and location where the clinician decides and computer-based delivery (Kawamoto et al., 2005). The screening reminders the quality director wanted will appear as a list in the visit's opening screen and as orders ready to sign, not as interruptive pop-ups.

Element Four: Drug Safety Alerts

Options: all vendor drug interaction alerts, or a tiered set. Decision: interruptive alerts only for severe interactions, allergies with documented severe reactions and dose errors, with less serious interactions shown passively. Rationale: a review of 17 studies found that between roughly half and nearly all drug safety alerts were overridden, frequently with good reason, because many alerts have low specificity and disrupt workflow (van der Sijs et al., 2006). Every unnecessary alert teaches clinicians to click past the next one, including the one that matters. Override rates will be monitored and alerts with high override rates reviewed quarterly.

Element Five: Order Sets

Options: individual orders only, or standard order sets. Decision: 24 order sets for common conditions and visits, such as diabetes follow-up, prenatal intake and medication-assisted treatment, built from current guidelines and reviewed by specialty leads. Rationale: order sets speed common work and embed evidence, while individual orders remain available.

Element Six: Security Roles and Sensitive Records

Options: broad access for all clinical staff, or role-based access with segmentation. Decision: role-based access, with records from the substance use disorder program labeled, visible to the treating team and shared outside it only with documented patient consent. Access to the records of staff members who are also patients is flagged and audited. Rationale: federal rules for such records and the privacy principle of minimum necessary access.

Element Seven: Patient Portal

Options: a basic portal for results and messages, or a full portal with scheduling, forms, payments and proxy access. Decision: full portal in English and Spanish, with mobile access, self-scheduling for routine visits, proxy access for caregivers and results released automatically except for sensitive results that a clinician should discuss first. Rationale: portal activation was 22%, and the network's patients rely heavily on phones.

Element Eight: Registries and Reporting

Options: export data to the existing population health tool, or use the record's built-in registries. Decision: built-in registries for diabetes, hypertension, depression, prenatal care and cancer screening, with automated federal health center program reports. Rationale: retiring the separate tool removes a fragile interface and gives care teams daily lists instead of monthly reports.

Resolving the Opening Disagreement

The physician and the quality director were both satisfied by the same principle: fewer interruptions, better placed support. Screening reminders moved from pop-ups to a care gaps list on the visit's opening screen, with pended orders the physician can sign in one step and a standing order protocol that lets medical assistants order routine screenings before the clinician enters the room. The quality director gained more screenings completed; the physician gained fewer interruptions. The design group adopted this as a rule for all decision support: put the right action in front of the right person at the right step, and interrupt only for danger.

Designing for Medical Assistants and Front-Desk Staff

Design often focuses on physicians, but medical assistants, nurses and front-desk staff perform much of the work in a health center. The design group included two medical assistants and a front-desk lead, who shaped rooming templates, intake questionnaires on tablets and a registration screen that captures sliding-fee eligibility, preferred language and social needs in one pass instead of three.

Downtime Design

Every design includes a plan for when the system is unavailable: printed schedules, read-only access to recent records on a protected workstation at each clinic and paper forms matching the templates, with a process for entering data after recovery. Downtime design will be tested before go-live.

Testing the Design With Users

Before build is final, clinicians from each discipline will run 20 scripted scenarios in the test system, timing tasks such as a diabetes visit note and a medication refill and recording confusion or errors. Designs that take longer than the old system for common tasks will be revised.

Measuring the Design After Go-Live

The network will track note length and share of typed text, time spent in the record per visit and after hours, alert override rates, order set use, screening rates, portal activation and safety events related to the record.

Conclusion

The new record's value will depend on eight design decisions: one shared record, short notes, a few well-designed decision support tools, tiered drug alerts, order sets, role-based access with protection for sensitive records, a full portal and built-in registries. Research on note length, decision support features and alert overrides guided the choices, and testing with clinicians before go-live will check them.

References

Kawamoto, K., Houlihan, C. A., Balas, E. A., & Lobach, D. F. (2005). Improving clinical practice using clinical decision support systems: A systematic review of trials to identify features critical to success. BMJ, 330(7494), 765. https://doi.org/10.1136/bmj.38398.500764.8F

Rule, A., Bedrick, S., Chiang, M. F., & Hribar, M. R. (2021). Length and redundancy of outpatient progress notes across a decade at an academic medical center. JAMA Network Open, 4(7), e2115334. https://doi.org/10.1001/jamanetworkopen.2021.15334

van der Sijs, H., Aarts, J., Vulto, A., & Berg, M. (2006). Overriding of drug safety alerts in computerized physician order entry. Journal of the American Medical Informatics Association, 13(2), 138-147. https://doi.org/10.1197/jamia.M1809