• Avoiding waste: Failure to thrive edition

    I’ve often kidded on the square that most of my research is spite based. In other words, something pisses me off, and I go on a mission to prove it’s wrong. My first ever publication was a systematic review showing that conservative therapies didn’t work to help babies with GERD (which Bill Gardner discussed a few weeks ago).

    Another pet peeve of mine was the care of failure to thrive. (Pediatric residents all over the country are nodding their heads in agreement right now.) Some babies either don’t gain weight as quickly as we’d like, or they start to lose it. A shockingly large number of them get admitted to the hospital. They get tons of tests and lots of consults, costing thousands of dollars, and the vast majority of them turn out to be functionally normal; they just need to be fed better. What irked me was that so many of them seemed to be admitted over the weekend. Almost all of the tests, and all of the consults, that we’d order would be unavailable then. So it seemed like these babies would just sit. They’d have longer admissions, and they’d cause me a lot of effort and paperwork.

    At least, this was my theory. I couldn’t prove it. I tried all the way back in fellowship, but I couldn’t get a good data source. That changed recently.

    Some of the people I work with have gotten really good at using the Pediatric Hospital Information System (PHIS). It’s an administrative database that contains inpatient, emergency department, and ambulatory surgery data for forty-three non-profit, tertiary care pediatric hospitals with teaching services in the United States. So I talked my super-awesome fellow (and soon to be junior faculty member) into studying this. The results were published a few months ago, and I’m embarrassed not to have blogged about them before now. “Increased Length of Stay and Costs Associated With Weekend Admissions for Failure to Thrive“*:

    OBJECTIVE: To evaluate whether admission day of the week affects the length of stay (LOS) and health care costs for failure to thrive (FTT) admissions.

    METHODS: Administrative data were obtained for all children aged <2 years (N = 23 332) with a primary admission diagnosis of FTT from 2003-2011 from 42 freestanding US hospitals. Demographic characteristics, day of admission, LOS, costs per stay, number of discharge diagnoses, primary discharge diagnoses, primary procedure code, number of radiologic and laboratory units billed during admission were obtained for each admission. Linear regression and zero-truncated Poisson regression were used for analysis.

    Basically, we looked at length of stay and the cost of admissions for failure to thrive based on day of admission. What did we find? As I suspected, being admitted on the weekend was  significantly correlated with a longer LOS and a higher cost. This remained true even when the analysis was restricted to children who turned out to have nothing else wrong, as evidenced by both an admission and discharge diagnoses of failure to thrive. The number of procedures for children admitted on the weekend wasn’t any different than for children admitted on the weekdays – so that didn’t account for the cost difference. Here’s the chart that keeps me warm at night:


    Average LOS, with associated costs, per admission day of the week for patients with a primary admission diagnosis of FTT and a discharge diagnosis of FTT.

    The blue bars (and the left Y axis) are the average cost of a failure to thrive hospitalization by admission day. The red line (and the right Y axis) is the average length of stay by admission day. It would be hard to make a better case that the increased length of stay associated with a weekend admit isn’t the cause of the higher cost.

    Our conclusion:

    Planned FTT admissions should, when feasible, be scheduled for weekday admission to decrease both health care costs and LOS. Although in some situations, such as an unsafe home environment or a critically ill child, weekend admissions may be necessary, in most cases we believe savings from this simple adjustment could be recouped without a change in health outcomes.

    There are so few times in medicine that we can save money with almost no impact on quality. This is one of them. Failure to thrive admissions that include weekend stays are often a waste of money. We should stop doing so many of them.


    *This was one of the first publications of a group I spearheaded here at IU, called the Clinical Utilization and Resource Efficiency, or CURE, Group. Our mission is to use data and evidence to find areas of waste in pediatric care. Expect to see more of our work in the future!


    • One of the highlights of a lifetime working with physicians has been my time working with the neonatal physician group at the local women’s hospital. One cannot describe the degree of calm that pervades the group. I suppose it takes a unique personality to take responsibility for the care of such small creatures in distress. Which brings me to my comment. Dr. Carroll’s study found that weekend admissions resulted in more radiological studies and laboratory tests. As to the former, I suspect it’s one of the byproducts of nighthawk: the practice of relying on non-group (and often offsite) radiologists at night and on weekends. Sort of like advising the patient to take two aspirins and call me in the morning, except in this case it’s nighthawk ordering a few radiological exams that will be studied in the morning.

    • I don’t have time to dig it up, but Dr. Carroll blogged about the price tag we pay for not having after hours and weekend access to our pcp’s. This would also seem to be, possibly, related to this as well.

      A concerned mother gets the itch late Friday night or on the weekend and doesn’t have much in the way of choices, so they wind up in the ER…

      • Actually, this gets at something that’s been nagging me. FTT doesn’t seem to me like an emergent diagnosis. I don’t mean that to say that it shouldn’t make parents worried, I mean that it’s not a sudden and unexpected emergency. You don’t go from the 95th to the 1st percentile of weight by age overnight. I would imagine that parents have at least some discretion in when they admit their kid – as opposed to, say, a systemic infection, where you’d admit them immediately, whatever the day of the week. Indeed, the article alludes to this, saying that many admissions could be scheduled for the weekday.

        Parents do tend to work weekdays. A large number of the patients in the study had public insurance, so we could assume that a lot of the parents may have had sporadic jobs that might have them working weekends. Either way, I wonder if there is something that can be done about the timing. Perhaps clinicians could help families plan these admissions better. Certainly, in a world of accountable care contracts, the hospital could save some money by doing so. (In a pure FFS world, it doesn’t matter to the hospital and in fact it’s more revenue for the hospital.)

    • I guess that hospitals think that it is more “efficient” to save money by not providing these tests on the weekend. They do generate more income by keeping kids in hospital longer but that is a perverse incentive.
      Is it too much to ask to have hospitals and doctors to schedule work on a 24/7 schedule? People are sick and require diagnosis and treatment 24/7. Shouldn’t the hospitals also schedule their services to match? After all, many other industries (hospitality, airlines, etc.) provide 24/7 services and seem to do it efficiently.

    • Are there studies investigating why parents / caregivers initiated the medical care on weekends, rather than on a weekday? I wonder if employment commitments, and access to child care for other children, factor into the decision to access care on the weekend.

      Also, a recent editorial in the MJA* brought up the issue of ‘efficiency’ of care – that is, that a patient presenting to emergency is likely to get diagnostic testing done in a single admission, rather than have to make multiple appointments with multiple providers, possibly in multiple locations, plus reviews by the PCP. For parents with a new infant, I can understand how this process could be completely overwhelming- and make a hospital admission appear far easier and preferable.

      * General practice patients in the emergency department
      Gerard J FitzGerald and Ghasem Toloo, Med J Aust 2013; 198 (11): 573-574.

    • @Weiwen writes “I would imagine that parents have at least some discretion in when they admit their kid”.

      Perhaps I misunderstand, but I thought the decision to admit rests with medical professionals, not with the parents?

      If a child gets admitted, it suggests that at the point of presentation, a medical professional decided that inpatient care, rather than management by PCP through outpatient and community services, was the most appropriate. What other codes could be used for this type of admission, apart from FTT?

    • I haven’t read the entire study but I can tell you anecdotally that as a General Pediatrician the most common reason that an infant might be admitted to the hospital on Friday as opposed to every other day was because those were the infants who were having the worst “failure to gain” or even weight loss, usually secondary to poor feeding, and no one was comfortable having them at home with their caregivers unmonitored for an entire weekend. They often would have been seen daily in a clinic all week and then only admitted at the end when they’d failed to improve or worsened.
      Some of these babies have been teetering on the verge of starvation for weeks (sometimes months depending on when their last visit was) and while you can easily do daily weight checks in the clinic during the week, many practices aren’t going to be open to do that over the weekend. Some babies failure to thrive is less severe and their admission for various studies and controlled/monitored feeding and weighing can be scheduled, but others are just too precarious to lose track of for 72 hours. At the end they limited their conclusion to “Planned admissions when feasible” but in my experience the majority of these admissions are not necessarily planned and would assume that the Friday ones especially are composed of the least routine. No provider in my experience would admit an infant and family to the hospital on Friday to be miserable for the weekend to no purpose if they felt safe leaving them at home till the following week.