How a Database Audit Fixed a Healthcare Platform Serving 400+ Cities

A preventive healthcare provider operating across more than 400 cities and handling over 75,000 checkups a year had an application that had stopped keeping up with it. Screens were slow, the data underneath was inconsistent, and nobody could say with confidence what a given field was supposed to contain. OnPoint Insights audited the database, fixed the integrity problems, realigned the data model and optimised the queries that clinicians were waiting on.

400+ cities, 75,000 checkups a year

The scale the platform has to hold up at, which is why database performance here is a clinical issue rather than a technical one.

Orphan records and key mismatches resolved

Referential integrity enforced across the application, so the data it holds can actually be relied on.

Faster access to patient records

Query optimisation and indexing of frequently accessed columns cut response times at the point where someone was waiting on a record.

The problem

The application was slow because the database was never built for this size

An application that performs well at launch can fail at scale without a line of code changing. Queries that were fine against thousands of records crawl against millions, and a data model nobody documented drifts as each team adds to it. What the business sees is a slow screen. What is actually wrong sits underneath.

At this client’s scale, that slow screen has a cost measured in waiting patients and clinicians. The database had grown without agreed definitions or enforced relationships, so performance and reliability were failing together.

What was getting in the way

  • There were no clear data definitions or standards, which slowed development and made consistency impossible to maintain.
  • Orphan records and mismatches between foreign and primary keys created integrity problems across the application.
  • Inefficient queries slowed the application and delayed access to healthcare records and workflows.
  • Database configuration, security and performance practices needed review for reliability and compliance.

Our approach

Audit first, then fix in the right order

What we found

Performance and integrity turned out to be the same problem. Orphan records and mismatched keys make queries do work they should never have to, and nobody could reason about the model because none of it was written down. Tuning the indexes on their own would have bought a few weeks.

What we did

A data dictionary the teams can work from. Definitions, relationships and usage guidelines documented, so consistency no longer depends on who happens to remember.

Referential integrity enforced. Orphan records cleared and foreign to primary key mismatches resolved, which is the floor underneath any reliable application or report.

The data model realigned to the business. The application’s model audited and redesigned around how the business actually operates, with room to grow into.

Queries, indexes and configuration tuned. Data types corrected and frequently accessed columns indexed, alongside a review of database configuration and security practices.

The choice we made

We fixed the structure before touching the indexes. Indexing a broken model is the quickest way to a demo that looks better and an application that fails again in six months. It meant the early weeks produced documentation and integrity fixes rather than visible speed, which is a harder thing to show a client and the only order that actually holds.

BUILT ON
  • MySQL Workbench MySQL Workbench
  • phpMyAdmin phpMyAdmin

The results

What changed for the platform

The application got faster, and the data underneath it became something the business could build on.

Clinicians wait less for records

Query response times improved after the indexing and data type work, which at this scale means faster access to a patient record at the moment someone needs it.

The data can be analysed, not just stored

With integrity enforced and the model documented, reporting and analysis on medical data became dependable, which is what any analytics work on this platform has to stand on.

Standards that outlast the engagement

The data dictionary and configuration improvements mean new development starts from written definitions rather than from someone’s memory of what a field means.

Let's Connect

Facing a similar challenge?

Talk to our data and AI team about where you are today and what good looks like. 30 minutes, no obligation.

Book a consultation
Professional customer service agent on a support call
More Case Studies
Next Case Study
Contact Us

Collaborate with us

We're here to answer your questions and help you find the right solution.

Client-oriented
Results-driven
Problem-solving
Transparent

"*" indicates required fields

This field is for validation purposes and should be left unchanged.