Database DevOps has come of age, as shown by a key finding from the survey data that 73% of organizations have already adopted Database DevOps across some or all of their projects or are planning to do so in the next two years.

In Healthcare, that’s old news, with 92% having adopted Database DevOps or having plans to do so. More importantly, 63% are already practicing Database DevOps, compared with 48% in other sectors, illustrating a forward-looking recognition of the advantages it brings.

Those in the Healthcare sector appear to be well versed in meeting the challenges, but the 29% who plan to adopt Database DevOps in the next two years still face potential obstacles. Across all sectors, the one major obstacle that stood out was a lack of appropriate skills in the team, with 42% citing it. In Healthcare, a different picture emerges and on every single count, the obstacles are far smaller.

As can be seen, the lack of appropriate skills is a concern, but it falls to 24%. Across all of the other measures, there is a similar fall, demonstrating that IT teams in Healthcare are far more knowledgeable and skilled in DevOps than other sectors.

The result of introducing Database DevOps can normally be measured by the faster time in which database changes can be deployed to production: the higher the adoption, the faster the frequency. In Healthcare, however, a more nuanced picture appears. 27% deploy changes in one business day or less compared with 31% across all sectors. Similarly, 43% deploy changes in one week or less compared with 53%.

It appears that for IT teams in Healthcare, the goal is different. Rather than having the ability to deploy changes faster, it’s being able to deploy changes on demand, when required, and do so safely and more reliably, with fewer failed deployments.

The type of data being stored and used in business has grown far beyond the structured data typically stored in relational databases. There’s now other data like semi-structured, unstructured, time-series, geospatial and graph data to contend with.

This in turn has resulted in businesses using more than one database platform, both on-premises and in the cloud, to collect, store and analyze the wide variety of data. Across all sectors, only 21% of organizations use one database platform, compared with 38% in 2020, for example.

Not in the Healthcare sector. Here, a vastly different picture emerges. The number of multiple database platforms is far higher, with only 7% using one platform, and 79% using more than five, compared with 29% across other sectors.

The most common databases in play remain the usual Big Four – SQL Server, MySQL, PostgreSQL and Oracle – but a wide variety of other database platforms are now also in use. These range from the embedded database engine, SQLite, used in medical and healthcare data storage systems, to NoSQL platforms like MongoDB, which can create a single view of patient records from a wide range of healthcare and medical data.

This dramatic shift in database usage was highlighted as far back as 2021 by the technology consultancy, Thoughtworks, in its fascinating article, Measuring the Value of Distributed Data Platforms in Healthcare. It similarly found that more than half of Healthcare organizations indicated they had five or more active data platforms in place to help in areas like delivering digital healthcare experiences and personalized medicine, understand changing patient needs and accelerating core payment processes.

This growing use of multiple database platforms raises, for the Healthcare sector in particular, important questions about how those databases are monitored for availability, performance and other issues.

As can be seen, the sector is far behind others in its use of alerts and notifications, real-time monitoring tools, performance metrics and every other concern that would normally be in place. This may be because Healthcare IT teams have built their own monitoring systems to cater for their particular needs which are different from other sectors. It may also be that regulations like HIPAA in the US or the GDPR in Europe prompt them to ensure high availability and data protection are their highest priorities.

With multiple database platforms in use, however, having an estate-wide view of every database, server, instance and VM becomes even more important. Firstly, to ensure that potential issues are identified as soon as they arise, wherever they are. Secondly, to have a common understanding across every platform of the most significant metrics.

Across all sectors the cloud has made a big impact and continues to do so. The top three reasons for doing so are the same for every sector with, interestingly, cost efficiency always coming in third place. In Healthcare, the importance placed in the ongoing robustness and reliability of database estates once again comes to the fore, with high availability moving into first place ahead of scalability:

• High availability and reliability
• Scalability and flexibility
• Cost efficiency

We can also see that the sector is treading a different path when it comes to how the cloud is being adopted alongside on-premises infrastructures.

Almost every organization in the sector has moved away from staying all on-premises and 46% now host their databases all or mostly in the cloud, compared with 36% across other sectors. This is in line with the higher adoption of multiple database platforms we saw earlier, many of which are cloud native. Note, however, the 24% which remain mostly on-premises is ahead of the 19% we see in other sectors.

When it comes to the future, this different shape for the cloud continues for the Healthcare sector. Across all sectors, a hybrid approach is favored by just over a third (36%), a full migration to the cloud by 19%, and 15% have no migration plans.

In Healthcare, four clearer options have emerged, each of which is similarly favored, with a hybrid approach slightly ahead at 27%. While nearly a quarter (23%) are still evaluating options, 21% are aiming for a full migration and 24% have no migration plans.

The picture that emerges is one where the players in Healthcare appear to have moved early. Yes, there has been a bigger move to the cloud, probably driven by the many different database platforms now in use for eHealth. There is still a requirement, however, for relational database platforms like SQL Server, MySQL, PostgreSQL and Oracle to handle the big relational database patient record and billing systems that remain in use. This may explain the quarter with no plans to migrate their systems from a mostly on-premises position.

Artificial Intelligence (AI) has arrived and in every sector is making a difference in the way code, content and images are being created. It’s the same when it comes to database management, with organizations of all sizes across all sectors using it or planning to do so.

It’s notable here that those in the Healthcare sector have been faster in their take-up of AI, with nearly a third (29%) doing so, compared with a fifth (20%) across other sectors. As with other sectors, AI is being used in testing and development for a broad range of tasks from working with database schema to automating testing scenarios, and from analyzing and optimizing database queries to generating sample code or snippets.

While there are more IT teams in Healthcare using AI, the benefits of doing so are the same that other sectors have experienced.

• Streamlining tasks/efficiency
• Automation
• Enhanced security
• Standardization
• Optimized database performance

There is one significant difference here worth mentioning, however. Across all sectors, each of the benefits were seen as broadly equal in their importance. In Healthcare, the streamlining of tasks was seen as almost twice as important as all of the others, signaling a distinct desire to gain the maximum efficiencies from using AI.

Just as there are expectations about what advantages AI can bring, there are also concerns about its use. Questions remain about the accuracy and source of the data being used in Large Language Models (LLMs), along with the reliability of the content that is generated.

Across all sectors, the concerns voiced were broadly the same. The one big difference with Healthcare, perhaps unsurprisingly, is that Regulatory Compliance moved from the number five position to number three:

• Data security and privacy
• Accuracy and reliability
• Regulatory compliance
• Training/expertise
• Ethical concerns

During database development, developers like to test their proposed changes against a copy of the database that is realistic and truly reflects the size and referential integrity of the database in production. Only by doing so can they be assured that their changes won’t be breaking changes that result in failed deployments.

This often leads to full-size production backups being used which, along with weak data security and access controls, puts Personally Identifiable Information (PII) at risk. Not so in the Healthcare sector which should be used as the benchmark sector for others to follow.

Firstly, we can see that over three quarters (78%) of the data used in development and test environments is synthetic data, compared with less than half (47%) across other sectors. A further 8% use a subset of production data and those who are ‘Unsure’ falls to a negligible 2%. Only 11% use a full-size production backup compared to 43% in other sectors.

At first glance, 78% in Healthcare having no approach for handling sensitive data appears to be a problem, but we’ve already seen that synthetic data is used by 78% of respondents in the sector, negating the requirement for any further measures to be in place.

We can also see that the 20% in Healthcare who limit access to sensitive data to specific users, mask or de-identify it, or replace it with synthetic data at a later stage, almost exactly matches the 19% who use a full-size production backup or a subset of production data.

While the picture of those in Healthcare present is promising, it also provides a wake-up call to any organization or business in the sector that does not meet the same standard.