Do hospitalists boost costs? - a reality or fiction?

The Hospitalists community is shocked. Or at least I am.

Yesterday Drs. Kuo and Goodwin published a very well written study in Annals of Internal Medicine. It is in fact, an impressive study despite all its limitations. They evaluated a 5% sample of Medicare patients in 454 hospitals, comparing 36,871 patients treated by their Primary care physicians (PCP) with 21,254 patients treated by Hospitalists during a period of 2001 through 2006.

They found that although the length of stay was lower (0.64 days less) among patients treated by hospitalists, and that the inpatient charges were S282.00 lower, the Medicare costs in the 30 days after patients were discharged were in fact $332.00 higher. This translates in an additional 1.1 billion in Medicare costs per year based on the approximate 25% Medicare admission managed by hospitalists. (This is very dramatic if we take into account that the government almost shut down yesterday).

In addition, the authors found that patients cared for by hospitalists may have a decreased lenght of stay hypothesizing that this may be at expense of increased rate of discharge to skilled nursing facilities or nursing homes by almost 20% compared with primary care physicians.

As well, they found that patients cared for by hospitalist had an 18% increase in subsequent emergency department visits and 8% increased risk of readmissions.

The findings were confirmed with an impressive statistical analysis in the entire cohort study of 205,190 admissions in 4657 hospitals.

However, this study had multiple limitations including: only included patients with an identified PCP; patients admitted with medical (non-surgical) diagnoses; did not include patients cared by subspecialists who may as well be hospitalists; only studied patients with fee-for-servide Medicare coverage; it included a period of 2001-2006 which may not be representative of the current practices; they extrapolated the costs, based on total Medicare charges, but did not actually directly assessed the costs.

Their last conclusion was very interesting - given that hospitalists may have incentives based on a fixed prospective payment based on the severity of the admission (medical diagnosis-related group or DRG) shifting costs to the outpatient in a post-discharge fee-for-service model, this drive overall to increased Medicare costs. The authors proposed the increase in bundling of payments based on the episode of care to minimize these incentives. The latter translates to the current proposed model of Patient-centered Medical home in which bundled payments will cover both inpatient and outpatient care.

I'm appalled. I am a strong believer that hospitalists in fact decrease overall costs of healthcare - why - because we provide evidence based quality care. We focus on quality improvement, increased patient safety, improved patient outcomes; we are very critical of our performance and advocates of increased accountability and transparency. We advocate the model of increased quality and decreased costs.

However this incredibly large national database is in fact, proving otherwise. In addition it proves some quality metrics to be in the red numbers area - for instance, shows an increased number of ED visits and readmissions.

If we take the heart failure model, it has been shown that early outpatient follow-up decreases readmission rates; but this requires a proactive behavior from all the healthcare system (including the patients). This is a current national priority and at least in this population we most likely will be seeing soon data of current practices.

I can argue that in the period of 2001 to 2006 a flawed system with limited access to outpatient medical care may have contributed to these results, and that this has been changing with time, and at least from my personal experience, we ensure that our patients receive a soon follow up with a PCP to facilitate transitions of care.

Now, it may be true that hospitalists may discharge patients more to skilled nursing facilities - but this may be due to the increased attention to detail to the patient's ability to perform activities of daily life and provide self-care; we may have increased Physical Therapy evaluations to ensure safer transitions of care. For instance, in this cohort of patients, they averaged 77.5 years old - any patient that age who is acutely admitted to the hospital may in fact suffer from substantial deconditioning that may not facilitate a safe discharge to home.

In addition, this analysis focus on economic impact and do not include all inpatient quality metrics, such as core measures, current national patient safety goals practices, use of VTE prophylaxis, documentation practices, etc.

An additional thing that is very important to take into account is the famous cost-to-charge ratios analysis in the Medicare population as well as differences in reimbursement in both inpatient and outpatient settings. In this study, the authors state that they calculated the Medicare spending 30 days after discharge based on total charges; as well they state that charges reflect price setting rather than resource consumption and therefore may overestimate costs.

So, this is a very statistically impressive study, but that focus on economic outcomes and may not be in fact evaluating the quality of care provided by hospitalists. In addition, Hospital Medicine has gone a long road since 2006 to 2011. We are very well embarked in the journey of patient safety and quality. We are advocates of the formula improved quality/decreased costs.

I believe further analysis based on newer populations, other payors systems (private insurances, Kaiser system, etc.), including more recent data, as well as more inpatient quality metrics, as well as adjusted severity and mortality should be done.

We need to be critical of our own practices and look in detail what can we do to improve. As the healthcare system in the United States moves toward the Patient-centered Medical home with bundled payments, most likely this presumable differences in healthcare expenditure between the PCP and the hospitalists model may in fact narrow to a non-significant level.

For the time being, the hospitalists will need to take this as an opportunity for reflection, and as a source of energy to fuel new research endeavors to improve patient outcomes and decrease healthcare costs.

For us as hospitalists our principle will always be....primum non nocere.