| Course | HI580 Information Systems Design and Implementation |
|---|---|
| Unit | Unit 4 |
| Paper type | System selection matrix |
| Length | About 870 words, 6 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | MS in Health Informatics |
| Updated | October 2026 |
Free sample paper for HI580 Unit 4
Weights Before Candidates: A Unit 4 System Selection Matrix Tested Against a Thousand Shifted Weightings
[Student Name]
Purdue University Global
HI580: Information Systems Design and Implementation
Unit 4 System Selection Matrix
[Instructor Name]
[Date]
Brackenridge Medical Center, its raters, candidates and every score are composite; the sources cited are real.
Criteria and Weights
Nine criteria were drawn from the requirement groups in Unit 2. Six raters, a labor nurse, a charge nurse, an obstetrician, the neonatal nurse lead, the risk manager and the interface analyst, each allocated 100 points across the criteria on March 4, before any vendor demonstration. Forcing a total of 100 makes raters trade one criterion against another rather than rating everything as important, which is the main weakness of simple rating scales in multi-criteria decisions (Triantaphyllou, 2000).
Averaged weights:
1. Newborn linking: 16.0.
2. Downtime behavior: 15.0.
3. Surveillance and archive: 13.3.
4. Bedside documentation: 12.7.
5. Retrieval for risk management: 10.2.
6. Five-year cost band: 10.0.
7. Interface burden: 8.0.
8. Vendor support: 7.5.
9. Training load: 7.3.
Newborn linking and downtime behavior carry the most weight, reflecting the twin registration errors and the network outage the unit remembers.
Candidates
Three approaches entered after the weights were fixed:
1. Candidate A: the record vendor's obstetric module, with surveillance built into the record.
2. Candidate B: a specialist perinatal surveillance system with its own archive, interfaced to the record.
3. Candidate C: keep the legacy central station and add bedside charting devices for the record.
Scores
Every cell was given a value between 1 and 5, drawing on vendor demonstrations, reference calls to two hospitals for each candidate and the interface analyst's estimate of build effort.
1. Newborn linking: A 5, B 4, C 2. Evidence: A creates the newborn in the same record natively; B links through an interface; C relies on the clerk.
2. Downtime behavior: A 3, B 5, C 3. Evidence: B's monitors store locally and replay; A loses surveillance display with the record.
3. Surveillance and archive: A 3, B 5, C 3. Evidence: B's archive met the retrieval demonstration; A's tracing review was slower.
4. Bedside documentation: A 4, B 4, C 2.
5. Retrieval for risk management: A 3, B 5, C 2.
6. Five-year cost band: A 4, B 3, C 5.
7. Interface burden: A 5, B 2, C 4. Evidence: B needs three exchanges and a shared clock.
8. Vendor support: A 4, B 4, C 2.
9. Training load: A 4, B 3, C 5.
Weighted totals, on the 1 to 5 scale: Candidate B 4.05, Candidate A 3.85, Candidate C 2.96.
The specialist system leads by two tenths of a point, a margin small enough that the paper does not trust it until the weights are shaken.
The Script
The script does three things only. It averages the six raters' allocations into weights. It computes each candidate's weighted total. It then runs 1,000 trials in which every weight is multiplied by a random factor between 0.75 and 1.25 and the weights are rescaled to sum to 100, recording which candidate leads each time. It makes no judgments about scores, which were entered by hand from the evidence above.
Sensitivity Result
Candidate B led in 992 of 1,000 trials and Candidate A in 8. Candidate C never led. The trials in which A led all combined a higher weight on interface burden with a lower weight on downtime behavior. The leader is therefore stable across reasonable disagreement among raters, but it depends on the unit continuing to value downtime behavior highly, which the raters confirmed in a follow-up meeting.
What the Scores Mean in Practice
The totals hide a trade that the unit should see clearly. Candidate A, the record vendor's module, is the safer choice for integration: one system, one clock, one patient identity, and newborn linking built in. Candidate B is the stronger clinical tool: better downtime behavior, a faster archive and better retrieval for risk management. The raters valued downtime and retrieval highly because the unit has lived through a network outage during a delivery and through a lawsuit in which a tracing took days to produce. Choosing B therefore means accepting more interface work in exchange for a system that keeps monitoring when the network does not. That is a defensible choice, but it moves risk from the bedside to the interface, and the project must fund that work rather than assume it.
Recommendation
Select Candidate B, the specialist perinatal surveillance system, with two conditions drawn from its lowest scores: the interface specification in Unit 6 must define all three exchanges and the shared clock before contract signature, and the training plan must account for the separate system. Selection research consistently warns that the costs of integration are underestimated at this stage (Kaplan & Harris-Salamone, 2009), which is why the condition is written into the recommendation.
Limits
The rater panel was small, six people, and two of them, the risk manager and interface analyst, weighted criteria very differently from the clinicians. Reference calls are anecdotal. Requirements practice suggests confirming high-priority requirements in a scripted demonstration before final contract (Wiegers & Beatty, 2013), and that step is scheduled.
References
Kaplan, B., & Harris-Salamone, K. D. (2009). Health IT success and failure: Recommendations from literature and an AMIA workshop. Journal of the American Medical Informatics Association, 16(3), 291-299. https://doi.org/10.1197/jamia.M2997
Triantaphyllou, E. (2000). Multi-criteria decision making methods: A comparative study. Springer.
Wiegers, K., & Beatty, J. (2013). Software requirements (3rd ed.). Microsoft Press.
What the HI580 Unit 4 assignment asks for
Unit 4 typically asks you to compare candidate systems or approaches against criteria derived from your requirements and recommend one. Expect a criteria table with weights, a scoring table with evidence, totals and a recommendation, sometimes with a sensitivity analysis. Six pages with an appendix is common. Fix criteria and weights before naming candidates, and say how the weights were set and by whom. Point allocation that forces trade-offs is stronger than rating each criterion independently. Give the evidence behind every score. If totals are close, test whether the ranking survives reasonable changes in the weights. Recommend with conditions drawn from the winner's weakest scores, and state the limits of the panel and evidence.
How the HI 580 Unit 4 example is put together
The criteria section lists nine criteria from the requirement groups and the six raters, explaining why forcing 100 points makes raters trade, with a multi-criteria decision reference. Averaged weights put newborn linking and downtime behavior first. Three candidates enter only afterward. Scores for each criterion carry evidence from demonstrations, reference calls and build estimates, and the totals follow; the highlighted sentence admits the narrow lead. A section describes the script's three tasks, and the sensitivity result shows the leader holding in 992 of 1,000 trials. A section on what the scores mean explains the trade between integration and clinical strength. The recommendation sets two conditions, and limits note the small panel.
Where the points sit in the HI580 Unit 4 rubric
Selection matrices are generally graded on the soundness of criteria and weights, quality of evidence behind scores, analytical rigor, the recommendation and presentation. The heaviest weighting tends to fall on how the weights were set, before any candidate appeared, and on and scores tied to a stated source. Rigor earns credit when close totals are tested rather than accepted, for example through sensitivity analysis. The recommendation is judged on whether it acknowledges the winner's weaknesses and sets conditions. Readable tables and a method a reader can repeat take the remaining marks. Matrices that assign weights after seeing candidates, give unexplained scores or declare a winner by a tiny margin without testing it tend to lose points.
Common HI580 Unit 4 mistakes, and how to avoid them
A common mistake in this unit is setting weights after the candidates are known, which invites bias. Another is scoring from impressions without citing demonstrations or references. Some students present close totals as decisive. Recommendations often ignore the winner's weaknesses. Code or spreadsheets are sometimes used without explaining what they do. Panels drawn from one profession skew the weights, so include at least one support role such as an interface analyst or risk manager. If your selection involves different candidates, such as cloud versus on-site archives or competing patient portals, list your criteria and raters for us, and the matrix will be built on that foundation. Sensitivity testing is also skipped more often than not.
Get an HI580 Unit 4 example written to your instructions
Whether or not your scenario names its candidates, send the Unit 4 prompt with the rubric and any requirements you produced earlier, and a custom matrix will fix weights first, score with sources and test how stable its winner is. The first sample is free, turned around in 24-48h. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More HI580 papers and MS in Health Informatics samples
- HI580 Unit 1: Discussion Board Post
- HI580 Unit 2: Stakeholder Requirements List
- HI580 Unit 3: Current Workflow Map
- HI580 Unit 5: Proposed Workflow Design
- HI580 Unit 6: Interface Specification
- HI580 Unit 7: Test Plan
- HI580 Unit 8: Seminar Reflection
- HI580 Unit 9: Go-live Plan
- HI580 Unit 10: Post-implementation Evaluation
- HI575 Unit 4: Risk Analysis Matrix
- HI501 Unit 4: Decision Support Rule Design
- HI570 Unit 4: E-Discovery Memo
- HI510 Unit 4: Ancillary Systems Review
HI580 Unit 4 questions, answered
Must the weights come from several stakeholders?
Not always, but a single author's weights invite the question of whose priorities they reflect. The example averages six raters and shows the spread between them, which reveals where roles disagree. If your scenario supplies no stakeholders, state the basis for each weight explicitly and consider testing how much the result depends on it.
What does perturbing the weights show?
Whether the winner depends on precise weights nobody could defend to a decimal point. The example nudges every weight at random within 20 percent across 1,000 runs and counts how often each candidate finishes first. A leader that wins nearly always is robust; one that wins three runs in four is a recommendation that carries a stated risk.
Should candidates be real products?
Many sections accept approaches described by type, as the example does with a vendor module, a standalone system and an upgrade. Where real products are required, cite published material with dates and avoid claims you cannot support. Either way, the scores should rest on evidence a reader can trace rather than on a vendor's reputation.
Where can I find a free HI580 Unit 4 sample paper?
The whole Unit 4 matrix is posted here, from point-allocation weights and evidence-based scores to the 1,000-trial sensitivity test.
Why test a selection matrix with shifted weights?
Because a small lead may depend on one set of weights. Rerunning the ranking under reasonable changes shows whether the leader holds when raters disagree.