Thursday, May 14, 2015
Comprehensive Care Management for COPD .... Kills?
It's not on-line yet, and the DMCB will link it when it becomes available. You can learn more about this chronic condition otherwise known as "COPD" here. And by the way, thusands of regular Disease Management Care Blog readers won't be that surprised by the study, because they had a warning about this back in March of 2011.
The Study Design
The authors set out to test whether a "comprehensive care management program" (which they abbreviated "CMMP") reduced hospitalizations or death among persons with COPD by recruiting patients with the condition at six Veterans Affairs (VA) hospitals. In order to be eligible for participation in the study, patients had to have been hospitalized for COPD in the preceding 12 months, be older than age 40, have breathing tests that showed significant airway disease and have a history of at least 10 pack years of tobacco exposure.
Patients were randomly assigned to either 1) CMMP or 2) usual care.
CMMP consisted of an educational program made up of four 90 minute one-on-one weekly educational sessions that included an overview of COPD, a review of the importance of self monitoring and instructions for self-treatment of exacerbations with antibiotics or prednisone (rationale here). The program was reinforced with a group session and monthly case management calls once a month for 3 months and then every 3 months thereafter. Patients got written "action plans," were encouraged to keep medications on hand and to call the case manager and or doctor when they initiated the antibiotics or the prednisone. There was a case manager at each of the participating sites.
The study was not double blinded, but the pulmonologists who reviewed the discharge summaries of the patients to determine the cause of the hospitalization or death were blinded to the treatment assignment group.
209 were assigned to CMMP, while 217 were assigned to usual care. Eight and ten patients dropped out of the CMMP and usual care groups, respectively.
The study was initiated in January 2007 and prematurely stopped in February of 2009. That's because 28 persons (17%) in the CMMP group had died, versus only 10 (7%) in the usual care group. Among the 28 patients, 10 deaths were ascribed to COPD, versus only 3 in the usual care group.
Patients were followed for another 6 months and there were an additional 15 deaths in the CMMP group, vs. 11 in the usual care group.
The difference in death rates didn't appear to be accounted for by any one of the study sites or any baseline factor, such as age, breathing test results or quality of life.
Up until that time, 97% of the CMMP patients had completed all four of the individual visits and 56% attended the group session. 89% of the scheduled case manager telephone calls were completed.
While there was a difference in death rates, there was no difference in hospitalization rates (27% and 24% CCMP and usual care respectively). There was no difference in the frequency of COPD exacerbations (around 4.4 per patient-year), but the CMMP patients were more likely to be treated with prednisone; there was no statistically significant difference in the use of antibiotics. While patients in the CMMP group indicated that they had greater self-confidence, there was no difference between the groups in quality of life.
When the DMCB reads the manuscript's "Discussion" as well as the accompanying Annals editorial, it's very clear that that the authors were very surprised by the results. They looked hard for a clinical or statistical anomaly and couldn't find one. As a result, this paper is either good science or just accidental fluke that runs contrary to the conclusions from multiple other positive COPD care management studies like this and this.
That being said, the DMCB notes this is a VA study that may not be generalizable to commercial settings. It also scrutinized the participant's baseline breathing tests and discovered that the mean FEV1 predicted was 38%, which is consistent with a severe level of disease. Assuming the results of the study are true, the DMCB wonders if CMMP would have worked for persons with more moderate COPD.
The results also speak to the necessity of constantly questioning cherished assumptions about the benefits of care management by subjecting it to repeated studies. While a randomized clinical trial is not always possible or necessary, comparing results to a valid baseline is well within the reach of just about any provider group, hospital, delivery system or ACO. This study teaches us that you never know what you may find.
Unfortunately, despite a critical mass of peer reviewed science that suggests care management "works," it's studies like this that can steal the limelight. Naysayers will need to be reminded that one single study like this should not be overinterpreted.
Last but not least, it's striking that results that were known about back in March of 2011 took a whole year to appear in print. Patients, their doctors and the care management service providers deserve better.
CODA: U.S News & World Report summary here.
Image from the NHLBI web site