New index quantifies risk of readmission or death after hospital discharge

TORONTO - Canadian researchers have developed a new index to score patients being discharged from hospital on their risk of dying or being readmitted in the next 30 days.

TORONTO - Canadian researchers have developed a new index to score patients being discharged from hospital on their risk of dying or being readmitted in the next 30 days.

If someone scores high on the so-called LACE index, then it could be a signal to physicians and others providing care that the patient will need extra attention in the days to come, says one of the researchers involved in the project.

That way, bad outcomes might be avoided.

"One of the things that one can do for a high-risk person is to follow them much more closely than you would otherwise," said Dr. Carl van Walraven of the Ottawa Hospital Research Institute.

"Lots of hospitals do post-discharge phone calls to see how people are doing, lots of physicians tell patients to give them a call if there are any problems ... hopefully that kind of attention to detail and followup will help avoid bad things happening after discharge."

Details on how the index was developed were published Monday by the Canadian Medical Association Journal.

Information was collected from almost 5,000 patients at 11 Ontario hospitals as they were leaving hospital, and the patients or their principal contacts were phoned 30 days later to see if there had been any urgent readmissions or deaths.

Of the 4,812 medical and surgical patients discharged, eight per cent died or were readmitted within the next month, the study showed. Of the eight per cent, 9.4 per cent died, and 90.6 per cent had an unplanned readmission.

Using the data, the researchers came up with a mathematical model allowing them to estimate the probability of a bad outcome, with four key variables emerging - the length of stay in hospital, acuity of the original admission, comorbidities and emergency department use.

"We were kind of surprised by it because we had 48 different variables for each patient, and some very detailed and we thought would be very predictive of outcomes post-discharge," van Walraven said in an interview from Ottawa.

"But after we ran the model we realized that we ended up with these four simple variables that could be very easily determined by a physician, but also very easily determined with administrative databases."

Then, using data from the Institute for Clinical Evaluative Sciences, the LACE model was externally validated in a million different patients in Ontario.

Van Walraven said the index is pretty simple for physicians to use, and they could have an application on a smartphone or PDA to help with the calculations.

"If you have a good information system within the hospital the index could be calculated automatically from data in the information system as long as you have a more advanced type of information system within the hospital," he added.

The moniker is derived from the four variables:

L - Length of stay: "The longer the person stayed in the hospital the greater the risk that they have a bad outcome post discharge."

A - Acuity of the admission: "If it was an acute admission rather than elective admission, there was more risk of bad things happening."

C - Comorbidity: "Basically a sum of all chronic diseases that person has."

E - "Number of emergency room visits that a person had in the six months prior to admission."

Van Walraven said another study is being started using the LACE index in another patient population, and he'd like to improve it.

"I think it's a good start, but the model itself is not as accurate as I'd like it to be, and so we want to try to see if there are other factors, other variables we can introduce to the model to try to improve its accuracy."

Dr. Norbert Goldfield, who has developed another evaluative tool called Potentially Preventable Readmissions, suggested that LACE developers might want to factor in that significant mental health and/or substance abuse disorders have higher rates of readmission.

In addition, he noted that longer length of stay might not mean a sicker patient every time.

"Longer length of stay certainly tends to correlate with severe illness, but it may also reflect hospital inefficiency," Goldfield said from Northampton, Mass., where he is medical director of 3M Health Information Systems.

"In addition, in a geographic area where primary care, home care and co-ordination of care is less than optimal, the number of emergency department visits tends to be higher than desirable, reflecting inadequate services rather than inherently sicker patients."

Goldfield wrote an accompanying editorial in the CMAJ.

Latest From ...
Most Popular From ...