MOST developed countries have undergone a revolution in data management in recent years. The availability of ‘big data’ has enabled entirely new avenues of research geared towards policymaking.

For example, detailed analysis of complete tax records has enabled public finance researchers to examine beha­vioural responses to tax-code changes in much finer detail than ever before.

Similarly, hospital exit records, insurance claims and demographic data have been linked to answer subtle questions of facility use, overuse, re-admittance and responsiveness to reimbursement rules. Information collected with an eye to its ultimate use and in a form that can actually address questions of policy relevance can also contribute towards setting the right policies.

This, however, requires systems to be in place to capture and use such data productively. This process is not automatic, and just ‘any’ data collection is not adequate.

Ten years back, a World Bank report discussed major information lacunae in Pakistan’s health and education sectors (among others where service delivery was an issue).

A decade later, while there is more ‘data’ available in one sense, in a deeper sense, we are no further along than we were then.

A recent study commissioned by the International Growth Centre (IGC) tried to determine the key factors impacting public health outcomes in Punjab.

Unfortunately, one was unable to analyse all contributing factors using the same dataset.

In fact, the analysis — based on two sets of population-based surveys, the Punjab Demographic and Health Surveys (DHS) of 2006 and 2012 and the Punjab Multiple Indicator Cluster Surveys (MICS) of 2008 and 2011 — was not able to provide answers to any serious questions concerning the correlation of policy inputs and health status.

As the Punjab government is geared up to expand primary healthcare provisioning around the clock, it really can’t be the case that there is no expectation that this will improve peoples’ health.

Some basic findings have emerged from the analysis. Economic status and mother’s education remain the most important variables in determining the health status of children (using heights, weights and mortality rates as indicators).

However, the impact of full immunisation on overall health status by these measures is undetectable, even though it is a good practice to have children fully immunised to protect against specific diseases.

These surveys also contain no information on water sources that could have helped determine their quality and, hence, their impact on health. ‘Piped’ water, the only proxy for quality, is definitely not the same as ‘safe’ water.

There is a lot of data that already ‘exists’ but not in forms that can be merged. School data is by school, clinic data by clinic, water data by watershed (or village) and so forth. Weak empirical results from the analysis can also be a consequence of an ineffective methodology, design and execution of the population-based/household surveys.

Hence, a high priority for the Punjab Bureau of Statistics, along with the main line ministries and departments, should be to work together to collect data that can be used for answering the simplest of questions concerning the basic determinants of health and general well-being of the province’s population.

One organising principle can be to create a ‘panel’, or a set of repeated observations over time on specific geographical units. This panel could enable ‘before’ and ‘after’ and ‘with’ and ‘without’ comparisons for any policy change that the government institutes without the need for special studies.

Such data can also provide a ‘rolling baseline’ for new policy initiatives. The long-run goal is to have a system of data collection that naturally increases in a form that becomes more and more useful over time.

A limitation on good data collection could be that some of the main actors in the policy arena may not really want to know the truth.

If we do not have an accurate measure of doctor absenteeism, representatives of doctors’ organisations can always deny the existence of the problem. If absenteeism is substantial, then they have a political problem. If it is not substantial, it only reinforces the position that they would claim anyway. What’s in it for them to find out?

Also, the full usefulness of the data will not become clear for some time and people may be unimpressed with its contribution in the meantime.

There have been piecemeal attempts to correct the situation with one-shot or uncoordinated attempts to fill in information gaps.

However, when these are done with little conviction, the resulting data remains unused.

So there is a vicious cycle of ‘data not being used’ leading to ‘little care for its collection’ leading to ‘poor quality data’ leading to ‘unconvincing conclusions concerning policy’ leading to ‘data not being used’.

The bottom-line is the fact that public primary healthcare has, to date, no discernible impact on the health status based on existing analysis. This needs to be explained before large sums of additional money are allocated to it. Getting more precise evidence on the role of public intervention on health is of utmost priority.

This article is based on a paper written by the authors for the International Growth Centre (Pakistan).

Published in Dawn, Business & Finance weekly, December 21st, 2015