In health care, not all patients are equal.

Some need costly treatment while others are relatively healthy, so providers must often decide who requires more attention and resources.

Reston-based information technology firm Altruista Health is attempting to automate that decision process, using its own predictive algorithm to identify a hospital’s sickest patients. Mining millions of health records, the 75-person company’s software performs a sort of triage, alerting physicians to patients statistically at risk of worsening and then providing treatment suggestions.

Altruista’s clients include hospitals and health-plan providers that pay annual subscription fees per patient, granting Altruista access to a trove of medical data, including health records, lab work and bills.

Depending on the client, the analysis varies. A health-plan provider might rely on the software to automatically identify higher-risk patients. A hospital might use it to flag a patient with multiple conditions — lung disease, congestive heart failure and depression, for instance — at risk of worsening. It might then delve deeper into the patient’s records, alerting the nursing staff that he or she lives with a smoker, which could aggravate the patient’s disease. The software might detect that the patient has not picked up a needed medication in weeks. Or it could make suggestions on alternative courses of care.

Altruista rolled out its platform in 2010 and was profitable by 2011. The firm declined to provide more specific financial data.

Its system is gaining traction. Eldorado, the health insurance division of Mphasis, an Indian IT firm owned by Hewlett-Packard, recently said it would offer Altruista’s technology as part of its offerings to health-care provider. Eldorado’s system helps manage cases for 5 million people and handles 35 million insurance claims each year.

“If they find a person with a particular condition, and they know they’re going to need a level of care, we can reach out to clinics and doctors and hospitals to negotiate particular deals or prices for that,” said David Hawkes, Eldorado executive vice president.

Affinity Health Plan, a New York nonprofit group offering health-care coverage to Medicaid patients, noticed a 50 percent reduction in hospital readmission rates in its first three months using Altruista, according to the company. Affinity serves almost 3,000 members.

Analysts expect steady growth in the health-care IT sector during the next few years as the industry adjusts to the health-care mandates pushed by President Obama.

“This is a golden opportunity for health-care [IT] firms to position themselves as partners to the health providers (hospitals, physician practices, etc.) by helping deliver cost-effective quality care,” Gartner analysts Jennifer Mazzucca and Anurag Gupta, who specialize in government and health care, said in a joint e-mail.

A slow start

Altruista got its start in 2006 when co-founder Ashish Kachru, then the national director of business risk management at the insurance giant United Health Group, realized hospitals could be better at deciding which patients needed care. At United, “we were really losing money based on what we were doing with management of patients,” he said.

While providers could eventually save money by preventing a patient’s condition from worsening, “we found out the average stay of patients on a health plan was about two years, and it took about two years to get them to an improved state of health before you’d start seeing any returns,” Kachru said.

So Kachru set out to make patient management easier. In partnership with the University of California at San Diego, his team built a program capable of harvesting and analyzing large volumes of medical data.

But Kachru chose a difficult time to found the company, just before the financial crisis took hold. Several promised venture investments never materialized, forcing Altruista to self-fund the project for about $4 million.

The bigger challenge was integrating Altruista’s software into hospital protocol, he said.

“That’s where the painstaking work was — understanding the workflow of each of these individuals, what they do on a day-in-and-day-out basis, and translating it into a platform that would make their work more efficient,” Kachru said.