If you fund or oversee hospitals
Numbers you can trace back to the rows that made them.
A figure nobody can drill into is a number to argue with rather than a question to answer. Every indicator here carries the cases behind it.
Quality indicators
Six numbers, each with its rows attached.
And one of them says not enough cases to say rather than printing a median of three. A district hospital's monthly numbers are small, and an indicator that pretends otherwise teaches people to distrust the whole set.
Every death already carries a cause and the clinician who certified it. The review queue is what turns that into something somebody is accountable for having looked at.
The monthly return
Built from the register, not retyped from a screen.
Every facility in the country sends a monthly return to the district health office. It is currently done by reading numbers off a screen and copying them onto a form, which is where the errors live and where a week of somebody's month goes.
In the shape DHIS2 reads
Generated from the same rows the clinic wrote, so the return and the record cannot disagree with each other.
Notifiable disease
Case notification raised from the clinical record, including the maternal death seam that crosses maternity and surveillance.
Programme funded care
Billed to the programme rather than to the patient, and claimable from it with the episodes attached.
The record
Tamper evident, not merely logged.
Every audit entry carries a hash of its own contents and of the entry before it. Editing a row out of band breaks the linkage, and deleting one breaks it too. The chain can be verified on demand rather than trusted.
Opening a patient record is itself an event with a name against it. Emergency access to a record somebody has no ordinary right to read is available, because at three in the morning it has to be, and every use of it is reviewed afterwards by a person.
Caseload
Never published on its own.
"How many cases is each clinician seeing" is the most misleading figure this product could publish. A doctor who saw sixty people in a day either had a light day of follow ups or an unsafe one, and the number alone cannot tell you which.
What kind of cases
New against follow up, and the acuity mix. Sixty follow ups and sixty first presentations are different days.
What else that clinician carried
Theatre lists, ward rounds and on call in the same period, because a caseload read without them is a caseload read wrongly.
And it flags rather than ranks
There is deliberately no composite score and no ordering by volume, however it might be labelled. A league table is how a quality tool becomes a stick.
Coverage
Who has fallen off, and is not coming back.
A defaulter list nobody works is a defaulter list nobody trusts. This one puts a name in front of a person, records whether the contact was attempted, and records whether it worked.
Consent to be contacted is asked separately from enrolment and held per channel, because a telephone number in a record is not permission to use it.
You can leave, and take everything.
A bulk export gives you every record at once, in a standard format, without asking us. It expires on a clock rather than only at the front door, because an export sitting on a server for a year is a copy of every patient record in the building.
We built the exit deliberately. A product that can only be left page by page is one you are not choosing to stay with.