While the debate rages around the travesty of justice in the Salman Khan hit and run case, the prolonged case has put a spotlight on the menace of drunk driving. Earlier in 2015, a Supreme Court Committee issued guidelines recommending jail term for drunk driving offenders. Also, with the 2015 holiday season, traffic police departments started scrambling to set up check-posts and plan other interventions to check drunk driving.

While action to crack down drunk driving is commendable, there is one crucial element missing — drunk driving statistics.

Glaring gaps in drunk driving statistics

India’s National Crime Records Bureau (NCRB) publishes annual data on traffic accidents and deaths caused from accidents. For the first time, the Bureau put out the causes for traffic accidents in 2014. Contrary to popular belief, the data shows that drunk driving constitutes a minuscule percentage of deaths. Besides, this cause-of-accident data is opaque. It does not give a detailed city-wise break-up of deaths by causes, nor does it slice the data on the basis of age, other demographics, or vehicle type. All of these are important parameters for designing interventions.

This data is also full of inconsistencies and gaps. In Tamil Nadu, which tops road accidents with over 67,000 cases in 2014, only 587 of accidents are due to drunk driving (less than weather-related road accidents, which are recorded as 712). There were 53 deaths from these 587 accidents, with 47 deaths registered in the state capital of Chennai alone.

NCRB data shows that there were only 6 cases of drunk driving deaths in all of rural Tamil Nadu (outside of Chennai) one year. This likely is not the full picture.

State-wise data shows these glaring gaps everywhere. Of nearly 25,000 accidents reported in Rajasthan in 2014, less than 1% (only 189) were reported as being caused by drunk driving.

Difficult questions

Studies show that gaps in drunk driving statistics are primarily caused by recording errors and human intervention. NCRB data was derided this year after an NCRB data entry error recorded that Vijayawada had 335 deaths due to drunk driving, while Bangalore recorded 13, Delhi 26, and Mumbai 2 drunk driving deaths.

It is a well-kept secret that corrupt officials collude with families of accident victims to refrain from mentioning “drunk driving” as a cause of accident (since it prevents families from making insurance claims).

Moreover, these are just drunk driving statistics that lead to accidents or death. Another big chunk of data that is largely missing from the table of policy makers and development sector stakeholders is the number of people (caught) driving drunk. There is no credible data available in the public domain on drunk people driving on our streets. The only way to get this information is to sift through traffic department challans. Even then, there is no credibility — such as in hit-and-run cases, when offenders are caught or surrender later — and there is no way to prove liquor consumption.

In addition, gender bias hinders good data, since police are reluctant to subject women drivers to sobriety tests. In the first few months of this year, Delhi traffic police prosecuted over 5,000 people for drunk driving, out of which only 12 were women.

Without knowing who is drunk on our roads, what they are driving, and what other high-risk behaviors they indulge in (such as speeding or jumping lights), we cannot know what programs to design, how to tailor the programs, or how to effectively reduce accidents.

Collecting better drunk driving statistics

In India, collating traffic accident data is a behemoth task, yet happens in a rudimentary manner. Every police station records First Information Reports (FIRs) and, from those, they segregate road accident reports. NCRB collates this data from police stations across the country.

Traffic accident data is replete with errors of omission and commission. These errors stem from the motives and biases of the police that gather data or incompetence of people entering the data.

With road accident deaths crossing 140,000 last year, and an apparent political will to deal with this through a strong road safety bill, it is time to look at overhauling our data-gathering mechanism. Desperate times call for desperate and creative measures. The best way to speed the collection of road accident data is by using an innovative, real-time data collection tool. For instance, police stations could use mobile data collection tools to record accident-related deaths from daily/weekly FIRs. After culling relevant data from FIRs, such as location of accident, age of victims, cause of accident, vehicle involved, etc., the police stations could make this data available to the NCRB and other bodies. These bodies can then review the data and spot any gaps or inconsistencies (like the errors in drunk driving statistics published in this year’s NCRB data) or regather data where required.

Technology can curb corruption to a large extent. Police pickets could use a dynamic data-gathering tool wired up with breathalyzers, thereby removing human discretion in recording incidents of drunk driving. Additionally, CCTV cameras at pickets could capture incidents of misreporting, such as women drivers going scot-free, thus making enforcement officials more accountable. For better data collection on drunk driving accidents, trauma wards in government and private hospitals where accident victims are usually brought in could use a mobile data collection tool. For every road accident patient admitted, trauma wards could collect blood alcohol data, input it into the tool, and send the data to government agencies.

There could still be gaps in gathering drunk driving data, but technology-driven data collection will go a long way in plugging the issue of missing and error-prone drunk driving statistics.