Domestic abuse can be an embarrassing thing to admit to; doctors from around the world often struggle with obtaining the truth from their patients regarding possible domestic abuse.  They may go to a doctor’s appointment to check up on health symptoms, but never reveal to the doctor that they have been victim to domestic abuse.  The doctor is put in a compromising position.  He is supposed to treat his patients, but does not possess the ability to.

That problem may be close to answered; thanks to a new computer program that has learned to pick up on possible clues left on the medical reports of patients.  The computer program can be used to save lives in predicting possible domestic abuse.  Patients that are victims of domestic abuse may do anything to avoid the embarrassment of being detected.  They may take measures to lie to doctors about injuries caused by their partner, or even travel to a different hospital to avoid detection.

The computer program has learned to pick up on clues that even the trained eye of a doctor or nurse may overlook.  The result?  A computer that understands the empirical differences between the medical reports of normal patients and the reports filed by victims of domestic abuse.  This was made possible by Ben Reis, a doctor from Harvard Medical School, who used the computer to compare the medical records of 561,000 people over a period of 6 years in an unnamed US state.  Of these 561,000 people, 19,000 were known to have been victims to domestic abuse.

The computer program then produced an algorithm that focused mostly around common patterns of injuries and mental illnesses of victims.  A few of these indicators are listed below:

- The computer found that certain patterns of injuries pointed to the specific violence of domestic abuse.  A few of these visible physical injuries include: bruises, bite marks, fractures in the face, and laceration.

- It also found that headaches, chronic pain, sleep disorders, fatigue or other signs of stress-related illnesses

- The computer found that alcoholism was a big predictor in women, and that depression and mood disorders were predictors of domestic abuse in men.

The computer program was not without error, however.  It was inconsistent with its detection, which can be expected for a new, free running program.  At times the program would suspect a case up to 6 months before the doctors did.  At other times, however, it failed to detect a case of domestic abuse at all.  The program will be tweaked to obtain both accuracy and a better success rate.

Either way, the program provides doctors with a tool that can be used to detect domestic abuse early.  It gives the doctor the option to choose whether to mention it and recommend treatment or to keep quiet.  In a world where domestic violence is prevalent in every society, the computer program may give some hope to the efforts of preventing domestic abuse.