Capstone Concept Lab: From Health System Problem to Informatics Project

Capstone Concept Lab: From Health System Problem to Informatics Project

Number of replies: 11

Most of you have already begun thinking about or developing an area or problem you would like to explore for your capstone. I have had several constructive conversations with many of you in order to help refine your capstone projects. I created this forum as an opportunity to help you put that idea into words, examine it through an informatics lens, and begin refining it with your colleagues, mentors, and faculty.

Let me mention that your idea does not need to be fully developed or approved at this stage.

In as much detail as you can, tell us about your proposed capstone idea. Address the following:

  • What problem do you want to address? Briefly describe the setting and why the problem matters.
  • How do you know the problem exists? What have you observed or experienced, or what preliminary evidence suggests that it is a problem?
  • What is the information or informatics problem? Where does information, data, workflow, communication, decision-making, or technology appear to break down?
  • What do you think could improve it? Describe your initial idea for an informatics-based intervention or improvement.
  • How might you know whether it worked? Identify one or two outcomes that could reasonably be measured.

End your post with a working capstone title. Don't worry about making the title perfect. I expect it to evolve as your project develops.

You are encouraged to read and comment on your colleagues' ideas. Ask questions, offer another perspective, or point out something they may want to consider. Help each other think through the problem, rather than designing the project for one another.

Important: Do not include patient-identifiable, employer-confidential, or proprietary information in your post.

Dr Chamberlain Nwanne

In reply to First post

Capstone Concept Lab: From Health System Problem to Informatics Project

by Salako Omolabake Fathiat -

Hi Dr. Chamberlain and Colleagues,

In high-volume suburban oncology clinic, we regularly care for patients who show strong signs of hereditary cancer syndromes like HBOC or Lynch syndrome. The core issue may not be a lack of clinical skill; it may be that our information pathway completely falls apart in these cases.

Right now, family histories are buried in unstructured paper notes or free-text EHR entries. Nurses and doctors have to recall complex NCCN risk criteria from memory during packed clinics. Worst of all, when we do refer a suspected high-risk patient, it is sent via an unmonitored email/phone call/paper hand-down with no tracking. We have no way of knowing if that patient ever got counselled or not.

The Informatics Problem & Solution:

I believe this is a structural breakdown in data capture, data decision and communication. To improve this without waiting for an entreprise-wide software change, I am working on a nurse-led, low-resource digital workflow prototype. It will feature: 

Structured Data Capture: Standardized fields for family history so the data becomes computable.

Automated Decision Support: Simple rule-based logic that automatically flags patients meeting NCCN criteria.

Closed-Loop Register: A status tracking system (from Flagged to Counselled or Tested) with a 48-hour exception list so no patient is forgotten.

How We’ll Know It Works:

I will be measuring whether the automated rule set catches high-risk patients accurately (targeting 95% sensitivity) and testing if 100% of flagged cases reach a clear, documented status in our tracking register.

Working Title:

Design and Validation of a Nurse-Led Structured Hereditary Cancer Risk Identification and Closed-Loop Referral Tracking Prototype in a Resource-Constrained Secondary Healthcare Setting

I would love to hear your thoughts, especially from anyone who has tackled referral black holes or closed-loop tracking in busy outpatient clinics!

Thank you very much.

In reply to Salako Omolabake Fathiat

Capstone Concept Lab: From Health System Problem to Informatics Project

by Khadijah Adefunke Jimoh -
Hi Omolabake,


You've identified a real problem and I love the solution you're proposing. I understand the implications of the missed risk identification because I'm working on something similar but it's not nurse-led. 

In reply to First post

Capstone Concept Lab: From Health System Problem to Informatics Project

by Ishola Ridwan Femi -

My proposed capstone focuses on improving the quality of malaria surveillance data in a hospital setting. The problem is that incomplete, inconsistent, or inaccurate data can affect the reliability of malaria reports and decision-making. I identified this issue through my experience with health information management and observing challenges with data documentation, validation, and reporting. From an informatics perspective, the problem involves gaps in data capture, standardization, and validation.

My initial intervention is to introduce a standardized malaria data capture and validation process, supported by clear data definitions and required data elements. I would assess its effectiveness by measuring data completeness and consistency before and after the intervention.

Working title: Improving the Quality of Malaria Surveillance Data Through an Informatics-Based Intervention.

In reply to Ishola Ridwan Femi

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Janiel Johnson -
Hi Ishola,

Your malaria surveillance data quality is an interesting topic. I like the before-and-after completeness and consistency measures. I'm working on standardized status definitions in a veteran tracking tool, so I can see why clear data definitions matter. Two things to consider: a small audit of recent records could show which fields are most often missing and strengthen your baseline. Also, involving the staff who enter the data when you design the required elements could help with adoption.
In reply to First post

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Olatunde Olaniyi Abiodun Oluwafemi -
Hello, Dr. Chamberlain and dear colleagues
My capstone proposal will focus on patient safety during nursing chemotherapy administration for breast cancer patients at Afe Babalola University, Ado-Ekiti, Multisystem Hospital (AMSH), Nigeria. A prior Failure Mode and Effects Analysis (FMEA) of the hospital’s breast oncology pathway identified chemotherapy administration as the highest-risk stage, with a Risk Priority Number (RPN) of 810. Key contributors to this risk include fragmented clinical data, limited availability of oncology trained nurses and specialists, inadequate communication and teamwork, weak monitoring and feedback mechanisms, the absence of standardized medication safety protocols, and frequent workflow interruptions.
A prior gap analysis of breast oncology workflow at the same center identified various challenges along the pathway. Rather than redesigning the entire breast oncology workflow, the proposal will target the chemotherapy administration step. The future redesign intervention is a structured nursing chemotherapy administration safety checklist integrated into the existing Electronic Health Record (EHR). The checklist will guide nurses through essential safety checks: patient identification; review of relevant documentation; confirmation of the prescribed regimen, dose, cycle, route, and timing; documentation of administration; and timely monitoring and recording of treatment-related side effects. It will include escalation prompts for discrepancies, omissions, delays, or adverse events. A designated nurse focal person will coordinate implementation, staff orientation, checklist completion monitoring, and communication of safety concerns to the oncology team.
Implementation will follow an eight-month, phased plan using the ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) change management model. Activities will include workflow validation, EHR configuration, user testing, staff training, pilot implementation, evaluation, refinement, and wider deployment. Success will be measured by a reduction in the chemotherapy administration RPN, high completion rates for the electronic checklist and critical verification fields, improved documentation quality, appropriate escalation of safety concerns, and acceptable user workload. Governance will have a designated nurse focal person under the supervision of the proposed multidisciplinary Breast Oncology Board, with monthly safety meetings, audit and feedback cycles, and structured corrective action tracking.
In reply to First post

Capstone Concept Lab: From Health System Problem to Informatics Project

by Dr Aminu Bello Liman -

Ahmadu Bello University Teaching Hospital, Zaria is a major referral centre for oncology cases from within Kaduna State and neighbouring states. Oncology department of the hospital coordinate the care of solid malignancies. Even though most patients access care on outpatient basis, hospital admissions are required in some instances. Oncology patients are admitted in various wards of the hospital based on the anatomical site of the cancer. For example, cervical cancer cases in Gynaecology ward, lung cancer cases in Cardiothoracic ward etc.

When required, oncology department Doctors are notified via written notes to review patients on admission. Due to the absence of single database, keeping record and monitoring of oncology in-patients is difficult as the patients are usually sparsely distributed across the hospital wards. 

To address this gap, our Capstone project is directed towards developing a cloud-based Microsoft excel spreadsheet that will serve as a tool for keeping record of Oncology in-patients admissions. 

To facilitate complete data capture, a Microsoft form linked to the excel spreadsheet was developed. QR code (linked with the form) is made available on the wards for the Oncology Resident Doctor to complete immediately after review of every newly admitted patient.

We hope this informatics intervention will serve as a dashboard to know the total Oncology admissions and their locations at a time. It will also provide a database which can be used for quality improvement audits and research regarding hospital admissions for cancer patients.

Some of the outcomes that will be used to assess success of this intervention are:

 1) Completeness of oncology admission records 

2) Ease of identification and retrieval of oncology admission information 

Title:

Development of Cloud-based Microsoft SharePoint Environment (Excel and Forms) in-patient information and tracking system for patients with cancer at ABUTH, Zaria, Nigeria.

In reply to First post

Capstone Concept Lab: From Health System Problem to Informatics Project

by Emelike Prince Samson -

My proposed capstone focuses on improving the management of cervical cancer screening information at the primary healthcare level in resource-constrained settings.

Problem and setting:
At PHC level, cervical cancer prevention involves identifying eligible women, documenting screening, recording results, and ensuring that those with abnormal findings are appropriately followed up. In settings where information may be spread across paper registers, individual records and spreadsheets, it can be difficult to maintain a clear picture of a patient’s screening pathway.

How I know the problem exists:
My experience working at a PHC and conducting my MPH research with 229 frontline PHC workers across 30 facilities exposed me to some of the challenges of collecting, documenting and managing health information. My experience with colorectal cancer symposium data also increased my interest in how cancer-related information can be organized to support better decisions.

The informatics problem:
The challenge is not simply having insufficient data. It is having data that are fragmented, inconsistently recorded or difficult to retrieve when needed. This can make it difficult to identify people who have been screened, determine their results, identify those requiring follow-up, and monitor what happens after an abnormal result.

Initial idea:
I would like to develop a simple digital prototype that organizes key cervical cancer screening data into a structured workflow. The prototype could capture eligibility, screening method, result, referral status and follow-up, with a simple dashboard to help PHC workers identify records requiring action. I am particularly interested in designing it around the realities of LMIC settings, where connectivity, equipment and digital literacy may be limited.

How I might know whether it worked:
I could assess data completeness and the time required to identify records requiring follow-up. I could also assess usability from the perspective of PHC workers.

Working capstone title:
“A Digital Data Management Prototype for Cervical Cancer Screening and Follow-Up at Primary Healthcare Level in a Resource-Constrained Setting.”

In reply to First post

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Janiel Johnson -

Problem

  • I'm a healthcare navigator in a VA-grant-funded program that connects unenrolled, high-risk veterans to VA healthcare. Many are homeless, low-income, and living with serious illness. When my caseload meets long VA wait times, veterans wait the longest and often don't know why. This matters because stalled enrollment can lead to preventable ER visits and mental health crises.

How I know it exists

  • I see it daily. Case status is mostly in my memory and notes, so I can't always provide an honest answer. I estimate fewer than 4 in 10 have a documented VA enrollment status within 30 days of intake, and VA eligibility responses take about 3 weeks. I plan to review the last 20 to 30 referrals to confirm this baseline.

The informatics problem

  • Information breaks down between my program and the VA. Our systems don't exchange data, case status relies on a single point of failure (me), and terms like "pending" and "lost to follow-up" aren't defined consistently, making handoffs difficult.

Proposed intervention

  • A shared Excel workbook, chosen for its low cost and familiarity, with a veteran tracker (standardized status and priority level), a referral log, a VA communication log that flags non-responses, and a dashboard showing who needs a call today. It would be stored securely, share only the minimum necessary information, and be built for VA partners to view or update.

Outcomes

  • Percent of veterans with a documented VA status within 30 days of intake, and median days to first VA response.
  • Percent of referrals completed within 30 days, and percent of updates logged directly by VA staff.
Working title: A Tracking and Communication Tool for Healthcare Navigation of Unenrolled, High-Risk Veterans
In reply to First post

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Purshottam Hoovayya -
Capstone Concept: Improving Oncology Information Management at AIIMS Bathinda

1. What problem do I want to address?
As a hospital administrator, I want to address fragmentation in oncology information across clinical documentation, pathology, radiology, tumour board discussions, and the cancer registry. Information is available across these areas, but compiling it for clinical decision-making, registry reporting, and research can be challenging. Incomplete or delayed information may affect coordination and data quality.

2. How do I know the problem exists?
From my experience in hospital administration, I have observed variations in clinical documentation and information management. Oncology information may need to be retrieved from different sources and compiled manually. I plan to assess the extent of these gaps through workflow mapping and a baseline audit.

3. What is the informatics problem?
The main issue is the absence of a standardised pathway for capturing, validating, and reusing oncology information. Diagnosis and staging details may be incomplete, tumour board decisions may not be documented consistently, and information may need to be entered more than once. These issues can affect data quality and make cancer registry reporting and research more difficult.

4. What could improve it?
I propose developing a Minimum Oncology Dataset (MOD) and a structured workflow linking clinical documentation, pathology, radiology, tumour board decisions, and cancer registry requirements. The initial work will focus on mapping the existing process, identifying information gaps, and developing a preliminary prototype. The aim is to capture information once, validate it, and reuse it wherever required, without replacing the existing hospital information system.

Since meaningful implementation may take longer than the capstone timeline, I will keep the initial scope realistic. I aim to complete the workflow map, baseline assessment, and prototype within the course period, with a limited pilot if time and institutional approvals permit. Full implementation and integration with existing systems can be pursued as a longer-term project.

5. How might I know whether it worked?
I plan to assess two primary indicators:
Diagnosis and staging completeness: Percentage of required fields completed.
Cancer registry abstraction time: Time required to retrieve and compile the required information for each case.

I will compare baseline and pilot results, where feasible, and seek feedback from clinicians, registry staff, and the IT team. Patient information will be de-identified wherever possible, and the necessary institutional permissions will be obtained before accessing or using patient-level data.

Working Capstone Title
Design and Pilot of an Integrated Oncology Information Workflow Linking Tumour Board and Cancer Registry at a Tertiary care teaching Hospital
In reply to First post

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Purshottam Hoovayya -
Capstone Concept: Improving Patient Safety Incident Reporting at AIIMS Bathinda

As a hospital administrator, I am interested in improving the way patient safety incidents are reported, documented, and analysed. I have previously used Google Forms for patient safety incident reporting, but the existing system has limitations related to the reporting process and user-friendliness. These challenges may affect the ease of reporting and the quality of information collected.

Since my proposed oncology information workflow project may take longer than expected due to the need for coordination across multiple departments, I have decided to begin with the patient safety reporting system as a more focused and practical informatics project.

My plan is to work with the IT team at my Hospital to redesign the existing reporting process and develop a more user-friendly software solution. The aim is to make reporting simpler, improve the completeness and consistency of incident data, and facilitate systematic review and analysis to support patient safety improvement.

I would begin by reviewing the limitations of the current Google Forms-based process, gathering feedback from users, and identifying the essential reporting requirements. The redesigned system would then be developed and tested on a limited scale, subject to technical feasibility and institutional approvals.

The project's success could be assessed through indicators such as the completeness of incident reports, ease of use reported by staff, and time required to submit a report. An important consideration will be to ensure that the system encourages timely, non-punitive reporting and supports organisational learning rather than merely collecting data.

Working Capstone Title: Redesigning Patient Safety Incident Reporting at a tertiary teaching Hospital: Developing a User-Friendly Digital Reporting and Learning System.
In reply to First post

Re: Capstone Concept Lab: From Health System Problem to Informatics Project

by Iwuanyanwu Anselm Chizurum -
This project addresses patient identity fragmentation and duplicate record creation in oncology care at a Nigerian tertiary hospital. Because cancer care is inherently longitudinal—spanning years and multiple departments like surgery, chemotherapy, and off-site radiotherapy—a fragmented identity poses a severe threat to data integrity, survival analysis, loss-to-follow-up tracking, and patient safety.

Fieldwork and on-site process mapping revealed that while a hospital-wide identifier (the hospital number) exists, nothing in the current registration or entry workflow verifies its use or performs a search before creating a new record. The only check performed at registration is asking the patient whether it is their first visit, which allows false first-visit claims and duplicate creation to go undetected. Furthermore, external evidence from South Africa's Provincial Health Data Centre and Nigeria's PEPFAR-funded National Data Repository demonstrates that systematic record linkage and deduplication are critical for resolving underlying data quality issues in similar settings.

The core information and informatics breakdown occurs at the point of patient entry and record creation, where registration requires no search against existing records and relies entirely on patient self-report. Information also breaks down across isolated diagnostic entry points—including Histopathology, Hematology, Oral and Maxillofacial Surgery, and Direct-to-Oncology referrals—and manual cancer registry abstractions, where records are compiled by hand across separate paper trails without automated cross-departmental matching.

To address these gaps, the proposed intervention introduces a connective, verification-focused matching layer that mandates a search-before-create step at every clinical entry point using deterministic keys like the hospital number and phone number, alongside probabilistic name and demographic matching as a fallback. To prevent unsafe auto-merges, uncertain matches are routed to structured human review, thereby formalizing the manual reconciliation process already performed informally by the hospital's cancer registry team.

The success of this intervention can be evaluated by measuring the precision and recall of the matching logic against a synthetic dataset constructed from confirmed fieldwork failure categories, such as false first-visit claims and spelling drift. Additionally, a directional estimate of the reduction in simulated duplicate record creation under the proposed mandatory search-before-create workflow can be compared against the current self-report baseline.

Patient Matching in Oncology: Building Reliable Longitudinal Data Across Fragmented Health Systems