00 · INTRODUCTION
Qualify takes a clinic from a patient call to a completed medication order: identity, eligibility, exams, pharmacy, payment and delivery, in one workflow.
01 · AT A GLANCE
One page, nine decision groups, one phone call to price an order. Nothing carried decisions forward.
Ask care type first. The flow branches to two clinic screens for an exam invite or seven for a medicated order.
Four findings from the live page, three with external sources, then six staff sessions that changed three design decisions.
Pricing moved on screen and required network pricing phone handoffs fell from one to zero by design. In testing, none of six reached for the phone.
02 · THE STARTING POINT
The original product put the entire journey on one page. Clinic, care type, pharmacy, package, patient details, payer, delivery, addresses and attachments ran down a single scroll that ended in two competing actions. It worked, and it asked the staff to hold the whole order in their head.
THE ORIGINAL WORKFLOW · 1 PAGE · 9 DECISION GROUPS
Good Faith Exams and Orders sits beside Prescription with Medication. Filling either one silently decides the whole order, and nothing on the page says which applies to this patient or what changes if the wrong one is chosen.
Change Your Custom Pharmacy
Pharmacy Package
Provider Selects (Best Price)
The chosen pharmacy renders as a data table down to NCPDPID, with no alternative to compare it against. Provider Selects, Best Price implies a comparison the interface never actually shows anyone.
Medication Delivery Method
Patient Home Address
Patient Shipping Address
Shipping Address
Home address, shipping address and same as home appear as separate controls, making it unclear what the clinic completes and what the patient provides.
Attachments
CHOOSE FILENo file chosen
Send Invite and Connect Instantly carry the same visual weight, with nothing explaining the difference or which one this patient needs.
03 · THE PROBLEM
What a clinic passes through to send one invite
Before · 1 long page · 9 decision groups · 1 required network pricing phone handoff
9 decision groups, no visible progress
After · 3 tracked stages · 2 care paths · 0 required network pricing phone handoffs
Good Faith: 2 clinic screens · Prescription: 7 clinic screens
Every decision carried forward
04 · PRODUCT AUDIT
I read the live page three ways: as a first time user meeting it cold, as the person who fills it every day, and as the order it eventually produces. Four findings recurred. The only quantitative claim I counted myself is nine decision groups on one page.
The only required network pricing phone handoff
INSIDE THE PRODUCT
OUTSIDE, ON THE PHONE
Pharmacy quoted by phone
Nothing saved when staff return
THE FOUR FINDINGS
Nothing on the page records what has been decided. With no running summary and no progress model, recall sits with the person filling the form, which is where entry errors get in.
SUPPORTING RESEARCH
Cowan’s model estimates a focus of roughly four chunks. This audit counted nine distinct decision groups on the live page.
Cowan · 2001 · Behavioral and Brain Sciences
Good Faith exams and medicated orders have different endings, but the page treated them as the same checkout.
SUPPORTING RESEARCH
In checkout research, form field count affects usability more than the number of steps. The same pressure applies when one form serves two outcomes rather than one. This is supporting pattern evidence, not healthcare specific validation.
Baymard Institute · 2024 · Checkout UX research
The page carried a Pharmacy Package field set to Provider Selects, Best Price, yet no price appeared anywhere on it. With nothing to compare on screen, the comparison can only happen on the phone. On the GLP-1 set, illustrative network quotes spread by $110.
ILLUSTRATIVE MARKET CONTEXT
Only 3% of surveyed patients reported comparing costs across providers before receiving care.
Mehrotra et al. · 2017 · Health Affairs
Three scattered checkboxes quietly decided whether the clinic or the patient handled pharmacy, address and identity. Nothing on the page states who owns the next step, so ownership is a side effect of defaults rather than a decision anyone records.
SUPPORTING PATTERN
When responsibility for a step is not stated on screen, it becomes a side effect of defaults. This is a general interface pattern, not a healthcare specific finding.
Pattern observation from this audit, no external source
05 · STRATEGY
A persistent order summary carries the patient, order set, price and logistics across every screen, so staff rarely have to reconstruct the order.
Answers finding A
The care type question moved from a buried field to the gate of step two. Everything after it exists only if the answer demands it.
Answers finding B
The pharmacy network became a searchable, filterable comparison. Eligibility filters, delivery and reliability rank, price breaks ties. Custom pharmacies sit outside it.
Answers finding C
Pharmacy choice, address entry and identification became explicit ownership choices between clinic and patient, replacing scattered checkboxes.
Answers finding D
06 · WORKING WITH OTHERS
The decisions involving prescribing, money and liability were worked through with the people responsible for that risk.
Signed off the prescribing boundary and what front desk may never release
Set the eligibility filters and confirmed what a quote can promise
Replaced live rate queries with a nightly refresh and an on-demand stock check
Required state licensing at step one instead of the review screen
Identified the top three support-ticket themes, all related to pricing
MY POSITION
Query every pharmacy live at the moment of comparison, so the price on screen is the price charged. Anything cached is a promise the product cannot keep.
ENGINEERING POSITION
Five live calls inside a form step adds seconds and a failure mode per pharmacy. Rate cards move slowly. A nightly pull plus one stock check is accurate enough and materially cheaper to operate.
WHAT WE SETTLED ON
Nightly rates, live stock on request, and a thirty minute quote lock. If the quote moves before send, the summary refuses to stay silent and the total never updates quietly.
DECISIONS I INFLUENCED BEYOND THE INTERFACE
07 · THE MAP OF THE FLOW
Care type is the branch that reroutes the whole journey: Good Faith ends with a sent invite, Prescription runs the full pipeline. Inside that pipeline a second branch, in network or not, decides how the pharmacy is found. Three switches assign ownership rather than route: who picks the pharmacy, who pays, and who completes delivery and identity. The final choice determines the ending: send an invite or connect live. This map covers the clinic half only. What follows, identity and consent confirmed, clinician approval, then release of the medication order, is described in sections 16 and 18.
Care type
Visit mode
Pharmacy choice
Who pays
Detail entry
Network availability
08 · WHAT EACH PATH REQUIRES
Splitting the paths only matters if the short path is genuinely shorter. These are counts taken from the flow, not timings.
Good Faith never reaches order sets, pharmacy comparison, payer or delivery. Those groups are not shortened, they are never asked.
The seven: patient information, care type, exam and order set selection, pharmacy, payment and delivery, review, send or connect. Sections 10 to 16 walk through them in order.
09 · UI FOUNDATION
Untitled UI provides the foundation: neutrals, corner radii, type scale and input patterns. Qualify adds the product voice through a periwinkle ramp built around its inherited accent. Every screen uses the same compact component kit.
10 · PATIENT INFORMATION
Step one captures identity and checks state eligibility while staff are still on the phone. The order rail begins with patient context, then carries each later decision forward.
11 · THE CARE TYPE FORK
One choice, two paths. Good Faith sends an exam invite after this step and skips pharmacy and shipping. Prescription continues through order set, pharmacy, payment and delivery.
12 · EXAM LIBRARY AND ORDER SET
Practice protocols live in one searchable library. Selecting one adds its order set and starting medication price to the summary rail. The final price is confirmed after pharmacy selection.
13 · PHARMACY AND PRICING
Six network pharmacies pass eligibility for this order set: state, accreditation, stock and cold chain. Search, price range and distance sort what remains, so network pricing is no longer a secret the clinic learns by dialling. A custom pharmacy sits outside this comparison and is still quoted directly.
14 · PAYMENT AND DELIVERY
Payment and delivery ownership are explicit choices. The clinic can enter the address and identification now or assign them to the patient, with the order held at review until completion.
15 · REVIEW AND SEND
Review brings identity, medical needs and logistics into one editable record. Each section links back to its source step before the flow continues to send or connect.
16 · CONNECT LIVE
The original page ended in two equally weighted buttons, Send Invite and Connect Instantly, with nothing explaining the difference. The redesign makes it a decision with consequences: send and the patient completes in their own time, connect and the patient is handed to a clinician for the live visit, order set and dosing already attached. Front desk prepares either ending. Only a clinician takes the call.
17 · VALIDATION
Six sessions across three clinics: three front-desk staff, two clinicians, one practice manager. Each ran one Good Faith invite and one medicated invite on the prototype. The goal was to identify breakdowns, not benchmark performance.
6 of 6
Completed a medicated invite
unaided, first attempt
1 of 6
Chose the wrong care path
noticed at the order set step, then corrected
5 of 6
Could not say who finishes the address
before the ownership line was added
0 of 6
Reached for the phone to price
the one handoff the audit flagged, unused
5 of 6
Used the summary rail to answer a question
without being prompted
WHAT HAPPENED
Four of six skimmed past the care type choice and only noticed at the order set step. The two options sat as equal weight radio cards with no consequence stated.
WHAT CHANGED
Each option now states what it commits you to before you pick. Good Faith says invite sent right after this step. Prescription says continues to order sets, pharmacy and shipping.
WHAT HAPPENED
Staff used the rail to answer questions but three of six still scrolled back up to re-check the pharmacy before sending, because the rail showed a name with no price attached.
WHAT CHANGED
The rail now carries the confirmed price next to the pharmacy, and marks anything still pending in the same line rather than in a separate state.
WHAT HAPPENED
Five of six could not say whether the clinic or the patient would complete the address, which is the exact failure the audit found on the live page.
WHAT CHANGED
Ownership moved out of checkboxes into a named line per item, clinic or patient, shown in the summary and repeated in the review screen.
18 · ROLES AND RESPONSIBILITY
The workflow moves between three parties: front desk, clinician and patient. Sections 10 to 16 name who owns each step. Section 16 is the one screen that shows a handoff happening, with a named clinician and a separate release action.
Every field change records the user, role and timestamp. This is a design rule agreed with compliance, not a screen in this study.
Checked at step one. An unlicensed state stops the flow before any clinical decision.
Patient uploads ID in the invite. The clinician confirms it before releasing the order.
Only a licensed clinician can sign or release an order. Front desk prepares, never releases.
Consent captured in the invite. Patient data stays in the record, never in email.
19 · THE PATIENT SIDE
Handing address, identity and payment to the patient only works if the patient half is shorter than the clinic half. It runs on the same three stages, so a stalled order is always stalled at a named step.
PATIENT SIDE PROTOTYPE FLOW
Name does not match ID
Held for clinic review. Never automatically rejected.
Card refused
Order stays intact. The patient retries without re-entering anything.
20 · STATES AND EDGE CASES
A flow is only as good as its worst day. These five states carry the rules that keep an order honest when the happy path breaks.
21 · PROTOTYPE OUTCOMES
The prototype holds the task together in one continuous flow. These are structural outcomes. Section 17 shows how they performed in six staff sessions.
