“Well, that’s the news from Lake Wobegon, where all the women are strong, all the men are good looking, and all the children are above average.”
Garrison Keillor, News from Lake Wobegon
As any reader of this blog probably knows, the average or mean is a single value that can be computed from any group of parameters with a numerical value. It is an accepted way to express what is typical of a group of values. Average and mean are synonyms not to be confused with the median, which is the middle value in a set of ranked values—the point where there are as many values below as above.
In many ways Judy Wientraub is anything but average. Diagnosed with ESRD at age 15, Judy has incorporated renal replacement therapy into her life for over 40 years. Recently she released a documentary, “Life on the Bridge…A Journey with Dialysis,” about her experience with all dialysis modalities including her current home hemodialysis treatment. The documentary title comes from the attitude of one of her early physicians who told her that dialysis treatment is just a bridge to renal transplant, which is the most desirable renal replacement therapy. Judy has a full, thriving life living on the dialysis bridge.
Judy does not sound like the average patient, but she insists that her dialysis experience is typical. She has an AVF that must be cannulated every treatment. She has dietary and fluid restrictions, renal bone disease issues, and frequent doctor visits. If we displayed her monthly lab or treatment data, it would be similar to many other thousands of hemodialysis patients. Her lab data is average, her problems are typical, but she is one of a kind.
During the Renal Physicians Association (RPA) meeting earlier this month I heard two data-related discussions that made me think about how we use “average”. At a workshop on Advance Directives in CKD the discussion started with the premise that everyone with CKD should have discussions about end-of-life wishes. In fact, probably all of us should make sure that we discuss our wishes with our family and should complete Living Wills as well as identify surrogate decision makers. During the RPA workshop the leader asked the audience, “What is the average lifespan for a dialysis patient?” The audience of nephrologists, renal social workers, and nephrology nurses offered the 4-5 year average that is often quoted. Someone in the audience also noted, “on average dialysis patient outcomes are worse than those for patients with colon cancer.”
Later that evening I attended the American Association of Kidney Patients (AAKP) Medal of Excellence Awards dinner in honor of people who have improved the lives of people with kidney disease. At my dinner table someone asked what is the longest anyone you know has lived on dialysis. I don’t provide care for Judy, but I mentioned her story and noted that she has had renal replacement therapy, including PD or hemodialysis, for over 40 years. Other nephrologists at the table had worked with people who had over 25-30 years on treatment. Caring for patients on dialysis over 10-15 years was not an uncommon experience.
The average dialysis patient lifespan is 4-5 years, but obviously we all know patients and have cared for many patients that are not average. Should we tell CKD patients about the average? Should they expect to have an average experience? If we treat or talk to them like they are average does that impact their expectations?
Not your average Joe
I like using data to improve clinical processes and to measure what is done and what is left undone, but statistics and analytics are definitely not my strong suit. I have to read a lot and talk to smart data colleagues to get by. One book I like is Naked Statistics by Charles Wheelan. In Chapter 2 “Descriptive Statistics”, Wheelan provides an example of how the average or mean is misleading. In this story there are 10 guys sitting in a bar drinking. Each guy has an annual income of $35,000, so the average income for the group is $35,000. Bill Gates walks into the bar and has a seat. Assuming Bill has an average annual income of $1 billion in round numbers, then the average or mean income of the group at the bar immediately becomes about $91 million.
In this example the math is correct, but the average does not provide a good description of the actual truth about the 10 guys at the bar. The 10 original guys at the bar are similar to each other, but they are not at all similar to Bill Gates.
As we focus on population metrics and use data like the “average” patient it will be important to make sure that we are describing similar groups of people. Even if a large patient group appears to have a normal distribution, subgroups within the population may not be represented by the overall average. All renal care providers surely recognize that older patients who start a trial of dialysis have a much different life expectancy than younger patients who need dialysis treatment to manage a chronic glomerular disease. These 2 patients may have similar lab work and they may require similar treatments, but their experience and outcomes will be vastly different.
“Hey, 98.6, it’s good to have you back again…”
Some characteristics of the dialysis population are well described by an average. For example, healthy humans have tightly regulated body temperature. The average oral temperature for healthy adults is 37.0o C (98.6o F) and thermoregulation keeps body temperature generally between 36.3o and 37.3o. In this case and for other tightly regulated homeostatic parameters such as serum potassium or sodium averages can be very meaningful.
The average is less meaningful when there is skewness in the distribution. Examples common in people with dialysis treatment are serum ferritin and PTH. The mean PTH for a dialysis population will be higher than normal, perhaps around 400 pg/ml, but there will be many patients in the 200-400 pg/ml range and a long tail of patients extending out to 1500 pg/ml or greater. Ferritin distribution in the dialysis group will range from very low, less than 100, to greater than 2000 with everything in between, not just a skew to one side of the mean.
While an average provides insight for homogeneous groups where the parameter is evenly and tightly distributed, the average life span of dialysis patients gives limited insight not only for individual patients, but also for the population as a whole. In the U.S. people receiving dialysis have great variability. The average lifespan does not apply to the active 90-year-old patient who might choose PD for quality of life for months or years that remain. The average is not applicable to the healthy 30-year-old patient with FSGS for whom the goal should be a normal lifespan using all renal replacement therapies necessary.
In reality, subpopulations are much more complex than even age and diagnosis. Big data and Watson-size computing power are needed to help identify subpopulations that are alike. As Terry Ketchersid describes in the February 15 blog post, using data fully and appropriately is important. Healthcare needs quality goals that represent appropriate outcomes for the proper patient groups, not just the “average” patient. We need informatics to illuminate expected treatment response and outcomes for unique patients, taking into account genetics as well as psychosocial and physical nuances.
At times it may be okay and even good to help people find out if they are an average part of a similar group of people. At other times we need data to reflect the individual experience. Judy Weintraub tells me that our healthcare attitude and language have often been discouraging to individual patients who do not belong to the average dialysis outcomes experience. We should be cautious that our average expectations are not taken to heart by people who do not realize that they are not average at all.
Dugan Maddux, MD, FACP, is the Vice President for CKD Initiatives for FMC-NA. Before her foray into the business side of medicine, Dr. Maddux spent 18 years practicing nephrology in Danville, Virginia. During this time, she and her husband, Dr. Frank Maddux, developed a nephrology-focused Electronic Health Record. She and Frank also developed Voice Expeditions, which features the Nephrology Oral History project, a collection of interviews of the early dialysis pioneers.