Real world data quality problems may involve many millions of records,and may require fuzzy matching of fields whose values (such as names) may only approximately agree. Fuzzy matching over large data sets (as needed, for instance, to match or de-duplicate patient records) is a hard computational problem, and at this stage many data quality tools run out of steam - and cannot  process the required numbers of records in the required times.

‍SmartSearch overcomes these limitations as it is able to scan many millions of records and find fuzzy matches. Flexible query strategies can be defined, to  support the assessment and resolution of data quality problems up to regional or national levels.

Some data quality issues can only be detected and corrected using approximate (‘fuzzy’) data matching – for instance to detect matches in the presence of spelling variations. This can be computationally very expensive when there are large numbers of records. SmartSearch can carry out fast fuzzy or Bayesian matching across millions of records – making the necessary data quality searches possible.

SmartSearch in England & Wales

With increasing integration between healthcare applications in England and Wales, high NHS number coverage in applications is needed for efficiency and patient safety. NovaMap Health has partnered with CareMotive to use their Spine Mini Services Provider for this purpose. Using SmartSearch to perform NHS Number lookups is the most cost-effective way to get high NHS number coverage, without expensive manual effort. There are several reasons for doing so:
‍
‍Legal Requirement: It is now a legal requirement in England to identify the patient by NHS number in all patient transfers between providers. You cannot do this if you do not know the NHS number for 20% or more of your patients.

‍Better Care Fund: One National Condition for the Better Care Fund - measured every 6 months - is the level of adoption of the NHS Number in each locality. As healthcare integration becomes more essential, this condition may be measured increasingly carefully, for all providers.

‍Clinical Adoption: Local integration initiatives allow clinicians to see the 'whole picture' for any patient across all providers. This is much valued. But if NHS numbers are missing for some patients from some applications, clinicians cannot see the whole picture for those patients. This may lead to slow adoption of local integration projects.

‍Patient Safety: As local integration of healthcare becomes the norm, clinicians will rely on seeing the whole picture for a patient across providers. Patients for whom they cannot see the whole picture will then be at risk. To run this risk for 20% or more of patients is unacceptable.

‍Data Quality: Data quality in healthcare applications has important implications for patient safety. Through CareMotive's SMSP, SmartSearch can retrieve many demographic fields besides the NHS number - such as GP practice details, and contact details. These can be used to improve the data quality of any healthcare application.

Other available mechanisms for NHS number lookup require high input data quality. Even small errors or variants in the 5 key demographics fields, such as patient name variants, result in failed searches. Correcting these involves expensive manual retries or data cleansing - costing many minutes or even hours per patient. Many providers cannot afford to do this. So their healthcare applications still have NHS number coverage levels of 80% or less.

If the first search fails, SmartSearch uses an Intelligent Automatic Retry strategy. - auto-correcting spelling errors and variants, and using wildcards. Then SmartSearch uses Advanced Data Quality techniques to check the quality of each match. This typically raises success rates from below 70% to around 95% - without any added manual effort.These results have been proven with several Local Authority social care departments.So either in batch mode or interactive mode, you can use SmartSearch to rapidly boost your NHS number coverage to high levels, and keep it there.