Commentary by Dr. Kevin Weingarten
Feature Article:Johnston, E. E., Bogetz, J., Saynina, O., Chamberlain, L. J., Bhatia, S., & Sanders, L. (2019). Disparities in Inpatient Intensity of End-of-Life Care for Complex Chronic Conditions. Pediatrics.
As a paediatric palliative care physician this article immediately caught my eye, as it deals with the largest portion of our consultation service – those with Complex Chronic Conditions (CCCs). In fact there are several articles this month relating to children and adolescents with serious chronic illnesses that are worth a closer look. However, what stood out for me about the article from Johnston et al, was their concern that sociodemographic disparities might be affecting our ability to deliver high-quality, goal-concordant end-of-life (EOL) care. As the article discusses in their introduction, families with children with CCCs already report unmet medical needs, lack of care coordination, poor provider communication, and psychological and social burdens of care. If, in addition to all of this, sociodemographic disparities are leading to less than ideal EOL care for these children, then we, as a PPC community, should take this article as a call to action call redouble our efforts for this vulnerable population.
The authors used an administrative database in California to study children with a CCC that died between 2000-2013. Specifically, they included 8654 residents of California who were between ages 1-21 at the time of death and had an ICD-9, or ICD 10th revision, code for a CCC either as 1) the cause of death on their death certificate or 2) the discharge diagnosis within 1 year of death. Using the administrative data base they were able to collect data on:
a) Sociodemographic variables including: Death, age, sex, race and/or ethnicity, payer, median household income in residence zip code, death year and distance from home to pediatric specialty center.
b) Inpatient EOL care including: whether there was a hospital death or any medically intense intervention(s) in last 30 days of life (eg. Hemodialysis, CPR, intubation or ICU admission).
For those of us in PPC the overall results weren’t that surprising. Two thirds (66%) of those that died did so in hospital. And 36% received medically intense therapy within 30 days of death. In fact, I would have predicted a higher percentage of children with CCCs would had received medically intense therapy near EOL.
In terms of the sociodemographic variables, the authors found that hospital deaths were more likely (Odds Ratio (OR) was approximately 1.2-1.3) if the patient was Hispanic, or “other”; and/or lived in a low income neighbourhood. They also found medically intense interventions were more common (OR of 1.2-1.3) for children residing in a low-income neighbourhood. Other factors with increased odds of a hospital death and medically intense interventions at EOL included: increased age at death (ie. 15-21 years of age), and distance from a pediatric hospital (>6 miles).
The published statistics paint a convincing picture that for those children with CCCs, sociodemographic factors impact their EOL experiences. However, I believe the authors, in their discussion section, have stretched their conclusions a little too far. I have chosen to comment on 4 of their statements from the discussion section, that I think actually distract us from what is otherwise a well-reasoned and important paper. And 1 truth that, fortunately, rings clearly through the distraction.
- “These sociodemographic variations raise concerns about whether the 20 000 US children that die annually of disease related causes are receiving quality EOL care that is consistent with patient and family goals…”. This statement is unnecessarily broad and beyond the scope of this article. This paper is not dealing with all child death in America, which is appropriate as focusing in on those patients with CCCs is essential. By using this quote to springboard to the population at large, the authors remove the focus on the very population they were studying.
- “In adult oncology, high medical intensity at end of life is considered poor-quality care, because medically intense EOL care is generally not goal concordant”. This is a referenced comment with good data to back it up. However, can we really compare the two populations of adults with cancer and children with CCCs? These two populations are very different (ie. diagnosis, treatments, prognosis, age, etc…). The authors even allude to this fact, in writing that “…children with CCCs have less predictable clinical trajectories”. In addition, they admit that “it is unknown whether families from low-income neighbourhoods desire more intense EOL care”. This uncertainty seems to break up their whole line of reasoning that more medical intensity at EOL = poorer-quality care.
- “Toxic stress associated with poverty may affect decision-making, particularly in high-stress situations.” This is also a true statement, yet is used in a way that places fault on the Substitute Decision Maker’s (SDM’s) for children with CCCs. In other words, the way I interpret this sentence in the context of the discussion, is as follows: A family is poor, so they have more stress, which negatively impacts their decision making, which leads to more intense EOL care, which leads to poorer-quality care. Meaning there is a direct link between a SDM’s decisions and poorer quality care. Now, some of this may be correct. I think that sociodemographic factors, particularly, low income, affects families at all stages of their child’s disease and especially at EOL. And so decisions are probably affected. However, as mentioned above in 2), we don’t know what these families want. So, it is entirely possible that higher intensity EOL care is goal concordant care. And therefore, while there might be toxic stress that we should evaluate and confront, it might not ultimately be negatively impacting the child’s EOL journey.
- “Nonetheless, all children with similar chronic conditions (same CCC category) should receive equally high-quality and high-intensity end-of-life care, regardless of sociodemographic status.” By now you may be wondering, what could be wrong with a statement like this? The short answer is: nothing. The slightly longer answer is that the paper is actually stating the conclusion of a future paper. What I mean by this, is that once future studies have actually shown that the quality of care is poorer – not just that there is increased intensity of said care – then we can conclude we have a problem. Right now, we only know we have differences in care, which I have learned are not always bad.
And so, with all of this being said, this article has nonetheless pointed out a really important truth, which, I am concerned, might get lost in their over-reaching discussion:
Non-white, poor children with CCCs get different EOL care.
For me, this is a really important fact that could carry the paper by itself. And as the authors mention, this difference, in it of itself, demands further research into: 1) patient & family preferences; 2) how EOL care in this population affects bereavement; and 3) what interventions can ensure goal concordant care for all patients with CCCs.
I believe this article is a call to action, but not because we are sure that those patients with CCCs and sociodemographic disparities receive poorer EOL care. Rather, it is a call to action precisely because we are not sure, and so, we must expand our understanding of this ever-vulnerable population.