Greetings. This is my first guest-post on Pallimed; I’m hoping it won’t be my last. I had envisioned a cleverer debut, elucidating my inclination to HPM vis-à-vis my bio, as a non-traditional (read mid-life-crisis-old) entrant to medicine, and my specialty (PMR by way of EM). However, Drew is taking a hiatus and the article I’m posting on now fell into my lap while still printing-press warm. So, thanks to Drew, Christian and to the entire Pallimed Editorial Board for the encouragement and invitation. On to Pallimed…

Drew just posted 'It just changes the complications' covering an ethnographic study that appeared in the Archives of Internal Medicine about the influence of nursing home (NH) culture on rates of tube feeding of patients with advanced dementia. My thoughts upon reading this (and I’ll wager I wasn’t alone) was that the information and the differences noted made sense to me, but it’s tough to make institutional and policy review and implement indicated changes based on ethnography.

So, as if in telepathic response to Pallimed readers, the current issue of JAMA contains an article about just this topic, abounding in hard data. Teno et al. 2010. Hospital Characteristics Associated With Feeding Tube Placement in Nursing Home Residents With Advanced Cognitive Impairment. JAMA. 303(6):544-550. The good people at GeriPal have also posted on this article.

This is a large, retrospective, chart-review covering the 8-year period from 2000 through 2007. It has a Texas-sized data set, N = 280,869 admissions for 163,022 residents of nursing homes with advanced cognitive impairment. The U.S. Nursing Home Minimum Data Set defined the study population. The resultant random sample size represented 20% of all Medicare beneficiaries who were previously non-tube fed NH residents age ≥66, admitted to 2797 out of 5401 U.S. hospitals during the study period.

The rates of endoscopic or surgical insertion of a gastrostomy tube (G-tube) placement during a hospitalization varied from 0 to 38.9 per 100 admissions (mean 6.5, median 5.3).

Encouragingly, the practice showed a decrease over time. The mean rate of feeding tube insertions per 100 admissions was 7.9 in 2000, decreasing to 6.2 in 2007, with the biggest decrease occurring in 2005.

I feared certain ugly truths possibly emerging that perhaps patients’ best interests would be sacrificed for medical student or resident training; or that hospitals with high specialist-to-PCP ratio would nudge patients down an unfortunate path. These fears of mine were not borne out by the data.

Hospital characteristics that did emerge as risk factors for G-tube placement in NH residents with advanced cognitive impairment were in order of decreasing adjusted odds ratios:

These differences persisted after controlling for patient characteristics. I think these represent at least unfortunate truths. To me there is a certain Kafkaesque darkness to the notion that admission to a large, for-profit hospital with high ICU use in the last 6 months of life results in significantly higher rates of G-tube placement in NH residents with advanced dementia.

Several nursing home resident characteristics were independently associated with G-tube placement. Black NH residents with severe dementia had about a 2-fold increased likelihood of having a G-tube placed; white residents had the lowest likelihood of G-tube placement.

Patient characteristics that reduced the likelihood of G-tube placement were having written advance directives, DNR orders, and orders to forgo artificial hydration.

Hospice use was weakly or not associated with feeding tube placement. This last factoid was perplexing to me. It led me to consider if this incongruity was a function of the heterogeneity and evolution of HPM practice models and styles over geography and time, e.g., placing venting G-tube placement for anticipated GI obstruction, or ex post facto, late hospice referrals.

Hypotheses anyone?