A quality-improvement initiative at University of Colorado Hospital has reduced dramatically the average time patients at low risk for tuberculosis (TB) spend in airborne isolation units to rule out the disease. The success comes at a time when the number of patients at risk for TB who are admitted to the hospital is growing.
The project, which involved representatives of several areas (see box), streamlined the process for ruling out TB in patients who have symptoms of the disease and must be isolated in negative air-pressure units. The components include an order set for physicians in the Epic electronic health record; an algorithm for nurses and respiratory therapists (RTs) responsible for collecting sputum for analysis; and a method for direct communication between the Microbiology Lab and Respiratory Therapy when a sputum sample is inadequate.
Work on the new policy began in April 2014 after a Department of Medicine Mortality and Morbidity conference, said Heidi Wald, MD, MPH, vice chair for Quality for the CU School of Medicine’s Department of Medicine. It launched in March 2015. The changes helped to reduce the average number of hours that patients at low risk for TB spent in isolation 38 percent – from a baseline of 87.6 hours to 54 in a five-month period spanning parts of 2014 and 2015.
Since 2013, UCH has had nearly three dozen confirmed cases of TB.
Mapping TB isolation
The success required a long period of data gathering and analysis. After mapping the previous work processes for assessing and isolating patients, the team found inefficiencies and inconsistencies. The result: longer-than-necessary stints in airborne isolation units that dissatisfied both patients and staff.
“We had no physician order set for the rule-out TB protocol,” Wald said. They set out to create a standardized process. Raven Astrom, RN, PhD, quality improvement clinical specialist, collected pre-intervention data to establish a baseline that she later compared to the post-intervention period.
Ruling out TB for a patient in isolation requires three sputum samples sent to the Microbiology Lab for AFB (acid-fast bacilli) smears. At least one of the samples must be collected around 6 a.m., when the patient’s first cough is “pure,” with minimal contamination from mouth bacteria, said Respiratory Therapy Supervisor Dana Wilkening, RT, who worked on the new policy. With three negative smears, a low-risk patient can be released from airborne isolation.
Sputum sticking points
The sputum-collection process sounds simple enough, but prior to work on the new policy, each of the three samples was collected at 6 a.m. That meant three days in isolation at a minimum. If the Microbiology Lab received a sample contaminated by too much bacteria from the patient’s mouth, the sample had to be re-collected.
Communication from the lab to the unit was frequently an issue in that situation, said Staci Aden, RN, associate nurse manager for the Pulmonary Unit, who was the unit’s clinical nurse educator when the project began. The lab called the bedside nurse to ask for another sample, Aden said. If the nurse couldn’t get a productive cough from the patient at the next collection time, he or she had to call an RT to induce sputum with a nebulizer. The hand-offs could lead to further delays, especially during busy days.
“The re-collection process became a sticking point that caused frustration,” Wilkening said. “If there wasn’t enough sputum collected, or if it was poor quality, the follow-up fell into a black hole.”
Meanwhile, the patient sat alone, waiting and wondering. “The next sample might not happen until days into the isolation,” Aden said. “Nobody wants to sit in an isolation room. Patients want a human connection other than seeing a provider in a big orange mask.”
The revised approach defines roles and sharpens communication. Physicians have an order set in the EHR that clearly explains the rule-out process with checkbox and radio-button choices; an HIV lab request was added last June. The intervals between the collections can be as little as four hours, as long as one is taken early in the morning. Nurses collect the sputum if the patient can cough productively. If not, the nurse contacts an RT for an induced sample. If the Microbiology Lab needs a new sample, the call goes directly to the lead RT, saving time.
Patients at moderate and high risk require blood tests and additional scrutiny to reach a diagnosis that rules out TB. The order set also prompts providers to contact Infection Control for a consult in these cases. Infection Control reviews the charts of every patient at risk for TB and has the final say on release from isolation, said Larissa Pisney, MD, associate director of Infection Prevention and Control at UCH.
An efficient rule-out process was also important because the number of patients at risk for TB that the hospital sees is growing, Wald noted. Data her team collected for respective five-month periods in 2014 and 2015, for example, showed the number of patients screened increased from 33 to 45.
Wald attributed that, in part, to a larger number of people in Aurora from countries with a relatively high incidence of TB. Broadly speaking, Census Bureau data for 2010-2013 shows that one person in five – 70,000 total – in Aurora was foreign-born, the largest percentage in the Denver metro area and more than double the percentage for the state.
As an academic medical center in the midst of a large urban community, the hospital also treats significant numbers of patients with histories of homelessness, IV drug use, HIV, prior exposures to people with TB, time spent in correctional facilities, and other factors that increase the risk of the disease.
Heidi Wald, MD, was sponsor of a quality-improvement project that cut the average number of hours patients at low risk of TB spent in airborne isolation units at UCH while protecting patients, providers, and the public health.
The TB risk at UCH is real, said Pisney. Since 2013, the hospital has had 34 confirmed cases of the infection, she said.
“We’re not a low-incidence facility,” Pisney said. “If we miss one TB diagnosis, it has huge public health implications as well as to the patient himself, our faculty and our staff.”
The hospital’s policy is to release patients from airborne isolation only when the likelihood of infection “is deemed negligible,” Pisney emphasized. She said she feels the new process has worked well to reduce inefficiencies while ensuring the hospital’s ability to contain TB exposures. She urged constant vigilance, noting that a patient’s clinical symptoms are only part of a portrait of TB risk.
“Because we have a lot of patients at risk for TB, it’s important that providers do a deep dive and take detailed social histories,” she said.
For Aden, the changes to the protocol highlight the importance for hospitals to define processes they follow.
“Any time we develop a more concrete way of taking care of patients, we are set up for success,” she said.
Airborne Isolation Project Team Members
- Sponsor: Heidi Wald, MD, Vice Chair for Quality, Department of Medicine (DOM)
- Clinical Champion: Larissa Pisney, MD – Infectious Disease
- Process Owner: Raven Astrom, RN, PhD, – Quality Improvement, DOM
- Core Team Members:
- Staci Aden, RN, Associate Nurse Manager, Pulmonary Unit
- Dana Wilkening, Supervisor, Respiratory Therapy
- Nancy Madinger, Director, Microbiology Lab
- Nick Walter, MD, Pulmonary Unit
- Jonathan Kurche, MD, PhD, Pulmonary Fellow
- Norihiro Yogo, MD, Infectious Diseases
- Extended Team Members:
- Sandy Godcharles, RN, Clinical Excellence and Patient Safety
- Jon Pell, MD, Epic Electronic Medical Record Physician Champion
- Brittany Cyriacks, RN, Clinical Nurse Educator, Pulmonary Unit