Migration Policy - D Brown - Submissions - 31 October 2007

Scrutiny Sub-Panel Scrutiny Office Morier House Halkett Place

St Helier

Jersey

JE1 1DD

31 October 2007

Population Register: Dear panel members

Having recently left the States but worked very closely with the design, creation and implementation of the draft Population Register I would like to provide the panel, and others, with further thoughts on why the island should be supporting the move to defining a single Population Register.

The States, like any large organisation, has grown over many years and have faced and continue to face many challenges in supporting an ever demanding and increasingly complex business. It also strives to support the citizens by providing efficient, cost effective services across a very wide range of disparate services from Health care to drain cleaning, from driving licences to managing tourism, from cradle to grave.

How many Databases ?

It is fair to say that almost every system has, at its core, a name/address database. Over many years the need to produce more systems quickly, to resolve a particular problem, has meant that yet another database joins the list of the many hundreds of databases that exist. All maintained by different people to varying degrees of accuracy and currency.

This is not just a States problem but a worldwide problem.

Each owner of each database is fiercely defensive of the quality of 'their' data. Each works in a silo and yet, in many cases, each interacts with other areas who have similar silos and so it perpetuates.

Data Quality

Databases consist of a number of related bits of information that a programme (and the user) needs to work, support, print, maintain etc. These 'bits' of information are called fields. They can either be fixed (ie have a value of 1 or 2 or 3 etc) or variable and be 'free' format.

As an ex system designer/developer I can advise that there is an inherent flaw with data fields that are free format, and that is that they ARE free format !. It is generally very difficult to enforce standards in this type of data unless you have a 'standard' to compare it against. So what happens ? Data is entered (in good faith) but because the address line will only hold say only 25 characters the input staff truncate the data,

so

La Grande Route de St Clement becomes

La Grnd Rte de St Clement

The human brain can easily see, translate and recognise the name but a computer would not generally see 'Grnd' , 'Grde' as 'Grande'

Colloquial names get used, 'Apartment' gets spelt with two p's and so becomes 'Appartment', and so on. The list of potential input problems is endless.

Accuracy and Currency

In the increasingly fast moving world of business, accuracy is one of the first things to suffer. The examples above are matched by others where names are spelled phonetically, the predominance of French street and house names are spelt with an interesting and often amusing range of accuracy/inaccuracy.

Then we move on to the names of people. Here the scope for variation increases dramatically. As well as the normal and expected typo, transposed letters and valiant attempts to spell long and generally complicated foreign names you get multiple entries for the same person with subtly different names.

The problem is that straight computer matching will see all these small differences and view them as different people perhaps living at different addresses.

Add to this the fact that the world does not stop and currency of data plays a huge part. Knowing I wrote to someone in the past 3 months at a given address provides very useful guidance to matching with an address that was last used say, 6 years ago.

Data Protection Principles

Any modern system developer worth his salt will ensure that any system design is fully compliant with the core Data Protection principles.

The 4th Data Protection principle states that 'Personal data shall be accurate and, where necessary, kept up to date'

Large organisations will try to be fully compliant with this principle but even the addition of one additional database is going to immediately undermine their ability to comply. Now add a few hundred of these databases and you start to see the scale of the problem. Indeed you could argue that by holding more than one database without any degree of automatic replication or synchronisation you have no chance of complying with this basic law of data accuracy and quality.

Central 'Single Point of Reference'

At the core of any resolution to this challenge is the basic design principle that maintaining something in one place provides a far greater possibility of ensuring accuracy and currency whilst minimising the effort to maintain it.

Some years ago when the States tackled the problem of a single property database there were many who were resistant to such change, each claiming their data was far better. By creating the single database (the Jersey Land and Property Index) and encouraging shared ownership and responsibility to maintain its accuracy and currency we have over 80 businesses and States departments now using this on a daily basis. Each time an inaccuracy or omission is spotted it is notified, viewed, validated and updated centrally. This ensures reduced maintenance, increased accuracy, timely updates, providing a single point of reference for all to use.

Turning to names. The same key principles in design apply. Single place, easier to maintain, less chance for duplication and replication of errors. The number of corrections that flow from an alignment exercise like this runs into tens of thousands. A data cleansing exercise from which everyone benefits.

It is also worth noting the fact that you may align a business database with a central database such as Property and /or Names but this does not enforce data sharing. It merely enables it should the departments have a justifiable and legitimate reason to do so.

So Why Align ?

As the foregoing tries to explain, many businesses (and its not just a States/Governmental problem) can benefit from using a single point of reference. It can ensure that high standards of data quality and currency are maintained. Fewer people are required to maintain and preserve the highest standards.

It can also underpin future improvements in technology by providing the ability for one area to communicate accurately with another area through means of a common reference point.

Having an increasingly accurate database of the islands population can only serve to support such key elements as:

Planning a timely and predictable property portfolio (linked to demographics not projections)

Providing the mechanism for GP out of Hours patient care and support

Predicting school growth in a timely manner based on location and current volumes

Planning other support services such as transport, infrastructure (power, water etc)

Creating support systems to manage civil emergencies with far greater accuracy

Once more States (and other) business areas have aligned and a formal consent strategy is in place the citizen will be able to feel that a more joined up approach is indeed possible, practical and safe.

Yours faithfully

David S Brown