Requirements Analysis for an
mHealth Application with
Midwives in Ghana
Olivia Vélez, RN, MS, MPH, PhD
Adjunct Associate R...
Centering Clinical Users in UserCentered Design (on a budget)
Olivia Vélez, RN, MS, MPH, PhD
Adjunct Associate Research Sc...
Overview
• What is requirements analysis?
• Why do we do it?
• Methods for requirements analysis

• Case study – Turning r...
What is Requirements Analysis?
• Determining stakeholder needs and translating them into
software/systems requirements
• W...
Steps for conducting requirements analysis
• Identify who you are building the system for (hint: might not be just the
end...
Methods for collecting requirements
• Focus groups
• Surveys
• One-on-one interviews

• Observations
• Contextual intervie...
What can we learn from focus groups/surveys/
one-on-one interviews
• What stakeholders think they need
• Issues
• Bounce i...
Why we need to do observations
Why we need to do contextual interviews
Methods for writing requirements
• Vignettes
• Use-cases
• Data flow diagrams
• User interfaces
Challenges of Health Information Systems in
Low-Resource Settings
• Health information system (HIS) implementations in dev...
Heek’s Design-Actuality Gaps Model
Project Context
• Millennium Villages Project (MVP)
• MGV-Net: Comprehensive open source e-health delivery
platform enabli...
Footer text is edited under "view/header and footer" menu

February 12, 2014

Page 15
Bonsaaso, Ghana
•
•

Population: 30,000
Primarily farmers and miners
Goals & Outputs
• Understand application domain to identify problems and opportunities
that can be addressed using mHealth...
Research questions
•

People
– What is the current workflow of MVP midwives?
– What are the roles and responsibilities of ...
Design & Procedures
• User centered design approach
• Participant Observation
• Contextual inquiry1
• Review of paper tool...
Turning data
sources into design
Goal
Is achieved by
enabling…
Use Case

Constraints: restrictions on
System What we aka o...
Participants
• Interviewed/observed 6 midwives in September 2010, 7 May 2011
Midwife

Interview 1

Interview 2

1
2
3
4
5
...
Results – People
• Midwives had heavy patient loads and intense work schedules
– Average 30 patients a day
– Every other w...
Monthly Reporting Requirements
Report Name

Key data elements

Addendum Antenatal/Maternity Monthly Data
Returns

ANC visi...
Poor Documentation Tools
Results – People continued
• Low technical self-efficacy
– “They have to come and train us so we are more confident with t...
Results - Organization
• Data sources are not searchable/easily referenced
• High potential for errors in current document...
Paper Record Storage

Medical Records

Personal Health
Records

Encounter
Registers
Fever Register

Action Taken
1. ACT & home
2. ACT & referred
Results – Organization Continued
• Only 2 full time technical staff members
• Distance to clinic will make supporting impl...
Results - Technology
• Power infrastructure is limited
– Clinics rely on solar power

• Network infrastructure inadequate
...
From Goals: Overview of Planned Forms
• Patient lookup and registration
• Capture patient register data needed for reporti...
System Qualities
System Quality Category

Accuracy
Documentation
Interoperability

Document analysis;
contextual inquiry;
...
Constraints
• Selection of OpenDataKit (ODK)
– Review of existing software available that met identified system qualities ...
Use Case Example
Use Case FR1. Enter new fever encounter
Primary Actor: MW
Preconditions: User was found or entered in pat...
Functional Requirements Example
Name
Gender
DOB
Age
DOB Estimated
NHIS #
Encounter date

Type
Text
Date
Number
Boolean
Num...
mClinic
mClinic
Current Status of mClinic
• Positive feedback from usability testing
• Existing deployments for capturing baseline immuniz...
Contact Information
• Email: olivia.velez@icfi.com
• Twitter: @mHealthNurse
• LinkedIn: www.linkedin.com/in/oliviavelez/
Acknowledgements
• National Institute for Nursing Research (P30NR010677)
• Health Services Resource Administration (1D11 H...
Olivia Velez -  Requirements Analysis for an mHealth Application with Midwives in Ghana
Olivia Velez -  Requirements Analysis for an mHealth Application with Midwives in Ghana
Olivia Velez -  Requirements Analysis for an mHealth Application with Midwives in Ghana
Olivia Velez -  Requirements Analysis for an mHealth Application with Midwives in Ghana
Olivia Velez -  Requirements Analysis for an mHealth Application with Midwives in Ghana
Upcoming SlideShare
Loading in …5
×

Olivia Velez - Requirements Analysis for an mHealth Application with Midwives in Ghana

1,586 views

Published on

Published in: Health & Medicine
  • Be the first to comment

  • Be the first to like this

Olivia Velez - Requirements Analysis for an mHealth Application with Midwives in Ghana

  1. 1. Requirements Analysis for an mHealth Application with Midwives in Ghana Olivia Vélez, RN, MS, MPH, PhD Adjunct Associate Research Scientist Department of Biomedical Informatics, Columbia University Senior mHealth Advisor Maternal Child Health Integrated Program(MCHIP)ICF International
  2. 2. Centering Clinical Users in UserCentered Design (on a budget) Olivia Vélez, RN, MS, MPH, PhD Adjunct Associate Research Scientist Department of Biomedical Informatics, Columbia University Senior mHealth Advisor Maternal Child Health Integrated Program(MCHIP)ICF International
  3. 3. Overview • What is requirements analysis? • Why do we do it? • Methods for requirements analysis • Case study – Turning requirements data into functional specs
  4. 4. What is Requirements Analysis? • Determining stakeholder needs and translating them into software/systems requirements • What do users need • What will you do to address those needs • Not how
  5. 5. Steps for conducting requirements analysis • Identify who you are building the system for (hint: might not be just the end-user) • Interview stakeholders • Write, review, and revise requirements with stakeholders • Write developer requirements and check for consistency
  6. 6. Methods for collecting requirements • Focus groups • Surveys • One-on-one interviews • Observations • Contextual interviewing
  7. 7. What can we learn from focus groups/surveys/ one-on-one interviews • What stakeholders think they need • Issues • Bounce ideas off of one another • Things they wouldn’t say otherwise
  8. 8. Why we need to do observations
  9. 9. Why we need to do contextual interviews
  10. 10. Methods for writing requirements • Vignettes • Use-cases • Data flow diagrams • User interfaces
  11. 11. Challenges of Health Information Systems in Low-Resource Settings • Health information system (HIS) implementations in developing countries have a failure rate of 20-25% (Heeks, 2006) • Causes of failure – Insufficient equipment/infrastructure – Poor project buy-in – Lack of resources to support intervention – Poor design and implementation planning • Disconnect between local needs and designers
  12. 12. Heek’s Design-Actuality Gaps Model
  13. 13. Project Context • Millennium Villages Project (MVP) • MGV-Net: Comprehensive open source e-health delivery platform enabling – Facility-based data capture – Community-based data capture of individual-level data – Data storage of patient health records – Automated mechanism for aggregating data and generating reports and feedback to healthcare providers and managers
  14. 14. Footer text is edited under "view/header and footer" menu February 12, 2014 Page 15
  15. 15. Bonsaaso, Ghana • • Population: 30,000 Primarily farmers and miners
  16. 16. Goals & Outputs • Understand application domain to identify problems and opportunities that can be addressed using mHealth • What is the future state we want to achieve? • Outputs: – Functional Requirements – System Qualities – Use Cases
  17. 17. Research questions • People – What is the current workflow of MVP midwives? – What are the roles and responsibilities of midwives at MVP facilities? – What is the current experience of the midwives with technology and their comfort level in learning new technology? – What are the information needs and information seeking behaviors of midwives working in MVP facilities in Ghana? • Organizations – What are the issues in collecting data from the health facilities? – What is the support capacity for and HIS implementation at MVP Ghana? • Technology – What is the current technology infrastructure at MVP Ghana? • Problems & Opportunities – What is the required functionality needed for the application based on the need and constraints of the application environment?
  18. 18. Design & Procedures • User centered design approach • Participant Observation • Contextual inquiry1 • Review of paper tools (document analysis) • Iterative Design through usability testing/evaluation • General interviews – Non-Governmental Organization (NGO) ICT eReadiness Self-Assessment Readiness Tool used to guide interview2 – Data quality assessment • Comparison to country and/or international standards 1. Coble, Maffitt, Orland, & Kahn, 1995 2. VanBelle, 2009
  19. 19. Turning data sources into design Goal Is achieved by enabling… Use Case Constraints: restrictions on System What we aka of Functional Description to user Use cases: Requirements: how GOALS:Qualities:want Non-what functional are system interaction such as and the requirementsto be achievesystem inputs, how does can implemented availability, task the system behave, etc. complete a security, how are the inputs processed, what are the Example: Appropriate testing outputs Data sources: Participant and treatment of all patients who interviews, interviews participant observation observation, Midwife present with a fever interviews, Data sources: use cases, contextual inquiry, document interviews, document and analysis, comparison analysis, Data sources: Midwifewith staff standards international standards interviews, contextual inquiry, document analysis, comparison with international standards Is achieved by enabling… Functional Requirements System Qualities Design Constraints …with these characteristics …with these restrictions Adapted from: http://tynerblain.com/blog/2006/01/04/foundation-series-structured-requirements/
  20. 20. Participants • Interviewed/observed 6 midwives in September 2010, 7 May 2011 Midwife Interview 1 Interview 2 1 2 3 4 5 6 7 8 X X X X X X X X X X X X X MW Years Exp. Nursing Exp. 20 Yes 26 Yes 3 Yes 43 Yes 4 No 42 Yes 8 Yes 2 Yes • Interviewed key staff members at MVP administrative site: IT manager, Health Manager, Data Analyst, Data Manager, Telemedicine Project Lead
  21. 21. Results – People • Midwives had heavy patient loads and intense work schedules – Average 30 patients a day – Every other weekend off, 24/7 call schedule – Little administrative support • Extensive reporting and documentation requirements – High degree of duplication
  22. 22. Monthly Reporting Requirements Report Name Key data elements Addendum Antenatal/Maternity Monthly Data Returns ANC visit information; Malaria prophylaxis; Delivery information Communicable disease surveillance report Malaria cases; Pneumonia cases, Diarrhea, AIDS, STDs Facility report of HIV test kit usage HIV test kits used Family planning returns Contraceptives administered Immunization and vaccine monthly returns Vaccines used and immunizations given by age and dose Institution monthly returns Malaria medication used Malaria reports of outpatient cases Malaria cases (by age and pregnancy) Malaria reports ITN/SP Stock Malaria medication used and stock holdings Midwives return ANC visit information; Delivery information; Postnatal data; Abortion data; PMTCT data Monthly data returns on ArtesunateAmodiaquine Malaria medication used; Children and pregnant women receiving treatment Medication and testing stock Malaria medication use; RDTs used National AIDS/STI control programme monthly returns HIV testing and treatment; STD testing and treatment; PMTCT data Outpatients Morbidity Causes of morbidity PMTCT monthly returns PMTCT data Returns on deliveries Delivery data Statement of Outpatients Patient age and insurance status Weekly notifiable diseases report Cholera, meningitis, measles, H1N1, Guinea worm, yellow fever, polio
  23. 23. Poor Documentation Tools
  24. 24. Results – People continued • Low technical self-efficacy – “They have to come and train us so we are more confident with the computer. We don’t know what we are doing with the computer. None of us.” • Limited information resources – Relied primarily on textbooks – No access to systematic reviews, journals, or other sources of up-to-date information • Perceived limited support
  25. 25. Results - Organization • Data sources are not searchable/easily referenced • High potential for errors in current documentation practice • Errors may go unnoticed for a long time
  26. 26. Paper Record Storage Medical Records Personal Health Records Encounter Registers
  27. 27. Fever Register Action Taken 1. ACT & home 2. ACT & referred
  28. 28. Results – Organization Continued • Only 2 full time technical staff members • Distance to clinic will make supporting implementation challenging • No infrastructure for remote support currently in place
  29. 29. Results - Technology • Power infrastructure is limited – Clinics rely on solar power • Network infrastructure inadequate – Inadequate signal strength at some of the clinics – Network outages – Internet outages at administrative site
  30. 30. From Goals: Overview of Planned Forms • Patient lookup and registration • Capture patient register data needed for reporting • Fevers (malaria) • Vaccinations • Prevention of mother-to-child transmission of HIV
  31. 31. System Qualities System Quality Category Accuracy Documentation Interoperability Document analysis; contextual inquiry; Rationale from data interviews with data A primary goal of this system is to improve the accuracy of data collection from the analyst/manager facilities. Data validation should be a key component of the interface Interviews with midwives; Easy-to-use, picture based manuals should be made available at the clinic due to the Interviews with IT lack to technical support available to midwives Staff Because OpenMRS will serve as the back-end the system must be fully compatible. Additional compatibility with other MVP mHealth and eHealth initiatives, particularly the telemedicine center is highly desirable. Learnability Due to the low technical self-efficacy of the end-users and the limited availability of technical support ease of learnability should take precedence over advanced functionality. Resource Utilization Midwives see about 30 patients in the morning. Hardware selection should support allow for this level of use without needing recharging Security The system will be used to collect patient data. The phones and the software itself should be secure. Remote deleting of data should be implemented in case the phone is lost. Data should be encrypted when sent over wireless network. Participant observation Participant observation; Interviews; Literature
  32. 32. Constraints • Selection of OpenDataKit (ODK) – Review of existing software available that met identified system qualities and constraints • Must work with OpenMRS • Must work on low-cost Android phones • Developed within contract requirements • Minimize text entry
  33. 33. Use Case Example Use Case FR1. Enter new fever encounter Primary Actor: MW Preconditions: User was found or entered in patient registration/Look-module and added to patient list Success end condition: Patient data is entered. System: Fever Register 1. MW selects patient from list 2. MW verifies patient demographic items 3. MW completes the following items Encounter date Temperature Duration of fever Test Done RDT or Malarial Smear results Danger signs of severe malaria Anti-malarial treatment given – – If RDT or Danger signs alert if no If not RDT and Dangers alert if yes If treatment yes, which medication Referral 4. MW uploads data to OpenMRS Extensions: 2a. Data needs to be updated, changes recorded to patient registration 4a. No network connection is available Data for use case development came from participant observation and contextual inquiry
  34. 34. Functional Requirements Example Name Gender DOB Age DOB Estimated NHIS # Encounter date Type Text Date Number Boolean Number Date Possible Values M/F DD/MM/YYYY Calculated from DOB Yes/No Alphanumeric 16 DD/MM/YYYY Duration of fever Number Number of days RDT results Danger signs, Malaria Treatment given Boolean Boolean Yes/No Yes/No Boolean Yes/No If yes, which medication Referral Text ACT, SP, Quinine, other Text Hospital, none, other Data Sources Contextual Inquiry (Document analysis) Participant Observation Data Analyst interviews Comparison to standards
  35. 35. mClinic
  36. 36. mClinic
  37. 37. Current Status of mClinic • Positive feedback from usability testing • Existing deployments for capturing baseline immunization data and verbal autopsy data by CHWs • Refining software and pre-implementation planning, seeking funding opportunities
  38. 38. Contact Information • Email: olivia.velez@icfi.com • Twitter: @mHealthNurse • LinkedIn: www.linkedin.com/in/oliviavelez/
  39. 39. Acknowledgements • National Institute for Nursing Research (P30NR010677) • Health Services Resource Administration (1D11 HP07346) • International Development Research Centre • Rockefeller Foundation • Novartis Fund for Sustainable Development • OpenROSA Consortium • Jonas Center for Nursing Excellence • National Library Medicine Biomedical Informatics Training Grant (5 T15 LM007079-20) • PAHO Collaborating Center at Columbia University

×