How not to conduct a survey

A simple questionnaire with a big flaw. Image from TSA Pre✓

by George Taniwaki

I recently received a voice mail message from the Transportation Security Administration. A woman’s voice told me that my Known Traveler Number (KTN) would be expiring soon and that I would need to renew it if I wanted to remain in the the TSA Pre✓ program. That’s the short line through security at the airport.

I haven’t been to the airport recently (and I hope you haven’t either) so I don’t know how long the lines are right now. But joining the TSA Pre✓ program is not expensive ($85 for 5 years) and has been worth it for me. So I pointed my browser to https://universalenroll.dhs.gov/ and started the renewal process.

Near the end of the process, I landed on a very unexpected page. It was a survey form asking questions about my flying habits (see screenshot at top of post). There are many problems with this survey that market research experts will immediately catch. But check out the fourth question. “How satisfied are you with your overall airport security experience?”

Geez, I hate airport security. It is intrusive, arbitrary, and time consuming. It also subjects you to radiation and chemicals of unknown safety. I guess it would be worthwhile if it effectively stopped violence and terrorism at a reasonable cost. Unfortunately, there is no evidence of efficacy and lots of evidence that it is really expensive.

Now, how should I answer this question? There is no explanation on the page about how your response data will be used. Specifically, there is no assurance that the responses will not be associated with your personally identifiable information (PII) and only aggregated data will be provided to the TSA.

Since TSA can make your life miserable, including revoking your KTN, the safest thing to do is to tell them you love your experience with airport security. Question 4 has 10 unlabeled radio buttons with the phrases “Extremely Poor” and “Extremely Satisfied” at the ends. I decide to pick the 9th button. High but not perfect. I figured anyone picking the 10th button will also be flagged for attention as either a liar or an obsequious bootlicker.

Anyway, as a marketer you may be tempted to increase response rate to your market research survey by integrating it into a customer transaction flow. Don’t do this. Your responses will be biased.

* * * *

Update1: Revised the third paragraph to clarify that there are many other problems with this survey. Thanks to my friend and colleague Carol Borthwick for reminding me that not all readers of my blog are survey experts. Below is a list of some of the obvious errors in this survey.

1. In the first question, how should one respond if you fly for both business and pleasure? And really, you fly to a destination for pleasure, you don’t fly because the experience itself is pleasurable. Almost nobody flies for pleasure, unless they are a pilot.
2. In the second question, what is the TSA trying to measure? My guess is the number of times respondents are screened by TSA in a year. A round trip usually involves two waits through the TSA line. However, one should not count trips on private aircraft where you don’t go through TSA lines or flights that originate outside the U.S., even if you go through U.S. immigration at the foreign airport.
   Further, if you have a connecting flight on a US domestic flight, you usually do not go through a TSA line again. If you arrive from an international flight and pass through immigration after the flight, you usually do go through TSA before boarding the next flight.
3. In any event, this survey was probably designed before the collapse in travel due to Covid-19. Does the TSA want to know the number of trips respondents took last year, this year (zero for me so far), or how many they would have taken if there was no pandemic. It doesn’t say.
4. What’s up with those weird ranges in question 2? And which radio button should respondents select if they fly exactly 31 times a year?
5. In the fourth question, notice that the wording of the two end point labels for the scale are not parallel. The low end should read “Extremely unsatisfied”. Also there are no labels for any of the intermediate points, leaving the distance between points up to the respondent’s imagination.

* * * *

Update2: Getting back to question 2, if you have a KTN, the TSA records each time you pass through security. So it should already know the actual distribution of how many times a year KTN holders pass through security. So what will it do with the survey data? Compare the response data to the actual data for accuracy? Check for lying and throw out outliers? Who knows.

 

Tracking the growth of Covid-19

Animated Covid-19 map, screenshot from Domo

by George Taniwaki

In order to make predictions about the future trajectory of the spread of Covid-19, you need to be able make sense of the currently available data. There are several steps to get good data.

Medical event data

First, you have to be able to collect data from multiple sources, clean them, and aggregate them based on a standard criteria. Each data record could include the following elements:

1. Event (what was counted, e.g., tests administered, positive test results, negative results, hospital admissions, ICU status, ventilation status, discharges, recoveries, deaths, etc.)
2. Location ID (where the event occurred, see below)
3. Date of incidence (when the event occurred)
4. Date of reporting (sometimes data is reported days or even months after the event and can be updated many times as errors are corrected or missing data is estimated)
5. Value (a count)

The best repository of Covid-19 data is maintained by the New York Times (on GitHub) with an interactive viewer. Johns Hopkins University Coronavirus Resource Center also has a dataset. The best source for counts of tests in the U.S. is available from the Covid Tracking Project sponsored by the Atlantic.

One of several graphics available from the New York Times

Public policy change data

In addition to medical events, there are public policy events that can be tracked, such as government orders to close nonessential businesses, travel restrictions, and so forth. These records could include the following elements:

1. Event (what type of public policy change was made)
2. Location ID (where the change applies to, see below)
3. Date of incidence (when the change was implemented)
4. Date of reporting (when change was reported, usually before the change is implemented)

Unfortunately, I could not find a centralized source of information on government restrictions and the dates they became effective. A different source of information that can help indicate how much contact there is between people is the amount of movement by people who carry smartphones. Smartphones contain a GPS antenna and can report their position. The position can be used to indicate what type of activity the person is engaging in. Google Health has a community mobility report that is updated regularly. An example report is shown below and the data in .csv format is available for download.

Among those who own Android smartphones and participate in tracking, trips have declined. Screenshot from Google Health

Demographic and geographic data

To analyze the data, you will want append demographic and geographic data about the locations. Unlike events, demographic and geographic data changes slowly, so only needs to be collected once during the model building process. The following data elements could be useful to prepare a model of forecast:

1. Location ID (from above)
2. Name or description
3. Location hierarchy (continent > country > region > state > county > city > zip code, etc.)
4. Latitude and longitude of centroid
5. Latitude and longitude of center of largest city
6. Surface area (km³)
7. Total population
8. Age distribution
9. Gender distribution
10. Income distribution
11. Race distribution
12. Political party affiliation distribution
13. Health insurance coverage distribution
14. Comorbidity distribution (smoking, diabetes, etc.)
15. Number of hospitals
16. Number of hospital beds
17. Number of ICU beds
18. Number of ventilators

Some good sources for this type of data are US Census, United Nations Demographic Year Book, United Nations Development Programme’s (UNDP) Human Development Report and the World Bank’s World Development Report, Gapminder, and ESRI.

Visualize the data

Once the data is aggregated, there are many ways to visualize it. Maps are an obvious way to display location data. Line charts are an obvious way to display time series data. Domo, a developer of business intelligence software, has very nice animation that displays time series data on a map (screenshot at top of blog).

Two caveats about their display. First, the number of cases is underreported because testing for infection was not widespread early in the pandemic, and is still too low today.

Second, outside the U.S. the data is by reported by country, not state or other smaller region. A single marker is used to represent the location of events. This is probably fine for Europe or Africa, where countries tend to be small. However, it is misleading for larger countries like Canada, Russia, China, Indonesia, Australia, and Brazil. Even data for a states like California is distorted because one would expect separate markers for the Bay Area and for the LA Basin instead of a single one in the middle of the state.

Johns Hopkins Center for Systems Science and Engineering has produced a nice dashboard hosted on ArcGIS (screenshot below). It does a better job of dividing large countries into smaller geographic partitions, but the colors are dark. A description of the project was published in Lancet Infect Dis (Feb 2020) and in a press release (Jan 2020). All of the data and the dashboard are available in a GitHub repository.

Another example of a Covid-19 map. Screenshot from ArcGIS

A note about line charts. You often see Covid-19 growth charts by country that display time (either calendar date, or days since the nth event occurred) on the horizontal axis and count on the vertical axis. Both are scaled linearly. I find these charts hard to interpret and compare. I think a better way to display growth data is to display data on the vertical axis using logarithm of counts per 100,000 population and on the horizontal axis using days since the n*(population/100,000)th event occurred. Even better would be to divide large countries into smaller regions so that all the charts covered regions with similar populations.

Making Forecasts

There are many groups making forecasting of Covid-19 infection rates and death rates. The CDC has a summary of them along with its own ensemble forecast. It predicts under 100,000 deaths in the U.S. at the end of May. The Institute of Health Metrics and Evaluation (IHME) predicts about 72,000 total deaths at the end of May but with a range from 60,000 to 115,000. You can download the data from the Global Health Data Exchange.

In addition to forecasting deaths, the IHME forecasts hospital utilization. These forecasts are used by hospitals to schedule resources and plan for peak usage.

Individual forecasts of cumulative reported deaths in U.S. from Covid-19 (left) and CDC ensemble forecast (right). Image from CDC

Cumulative death forecast in U.S. Image from IHME.

One of the best forecasts I have seen was produced by the Economist. It synthesizes data from US Census, New York Times, Covid Tracking Project, IHME, Google Health, and Unacast. The choropleth map of the U.S. below shows risk factors for Covid-19 mortality at the county level. Green shows areas where the risk level is low (less than 1%) and red shows high (6% or above).

Dixie in the crosshairs. Image from Economist

* * * *

Update1: In just one day, the IHME forecast is obsolete. See my response at https://realnumeracy.wordpress.com/2020/05/04/tracking-the-growth-of-covid-19-redux/

Update2: Add link to New York Times dataset and interactive viewer

 

The Covid-19 tracking app won’t work

Track this. Photo from Bloomberg BusinessWeek by Karen Ducey/Getty Images

by George Taniwaki

In a Bloomberg Businessweek editorial (Apr 2020), Cathy O’Neil (mathbabe) explains why a Covid-19 tracking app won’t work. It’s all about self-selection bias.

* * * *

Update: For a good non-technical description of how the Apple and Google contact tracing API works, including the encryption method, see Economist, Apr 2020. The article also suggests that even though using an app for contact tracing is imperfect, its low-cost and passive nature makes it worthwhile.

 

Kidney donor outcomes vary by race

As mentioned in a Nov 2009 blog post, there isn’t very much data on the long-term outcomes for live kidney donors. That’s because they are not being tracked. Further, there is little data on what attributes (independent variables) may indicate which donors (cases) are more likely to suffer adverse outcomes (dependent variables).

Harvey Mysel of Living Kidney Donors Network recently posted a link on Facebook to an article that shows that medical outcomes for living kidney donors vary by race. The study in the New Engl. J. Med. Aug 2010  (subscription required) caught my attention because two of the authors, Connie Davis and Paolo Salvalaggio, are at the Univ Washington Medical Center where my donor surgery will be performed. Dr. Davis is a nephrologist and director of the kidney transplant program. Dr. Salvalaggio is a surgeon in the program and was originally assigned to be the surgeon for my nephrectomy. (A schedule change led to a change in surgeon.)

They used a clever technique called a retrospective study to find the outcomes of donors. Rather than ask donors as they enter a transplant program to participate in a longitudinal study (called a prospective study) they looked at historical medical data after the fact. They obtained the historical medical data by matching the ID of donors in the United Network for Organ Sharing (UNOS) database with the customer database of a cooperating health insurer (the insurer is not identified, but my guess is Kaiser Permanente). Retrospective studies are fast (no need to wait several years to collect data) and inexpensive (no need to track patients for years as they move, stop cooperating, change insurance plans, etc.). However, these studies are subject to many types of sampling bias, which are beyond the scope of this blog post.

The authors make two findings. First is that some donors, both black and white, receive treatment for hypertension, diabetes mellitus, and chronic kidney disease after their surgery. Second is that black donors had higher prevalence of these morbidities than whites for all three conditions. On their own, these findings are not particularly surprising since these three diseases are very common chronic conditions and the black population as a whole has higher rates than whites.

However, it does lead to two concerns. The first is that although kidney donors are healthier than the population at large, doctors must not assume they will remain so. They should be vigilant for signs of chronic diseases among their patients who were kidney donors. This study shows that even within a few years someone who was thoroughly tested (and kidney donors get an extremely detailed examination) may begin to show symptoms of chronic disease. Hypertension, diabetes mellitus, and chronic kidney disease are often called silent killers. This study shows just how silent.

Second, the article says prevalence of these diseases among certain groups of kidney donors were in some cases as high as or higher than expected for a similar subpopulation that were not donors. This deserves additional research. Using prevalence rate (proportion who have the diagnosis) rather than incidence rate (proportion who receive their first diagnosis) may understate the seriousness of the problem. That’s because within the general population, the prevalence of these three chronic conditions is higher than it was for the kidney donors during the year in which they underwent their donor surgery. Thus, if the prevalence of these chronic conditions is the same as the general population in later years, then the incidence rate each year among kidney donors must be higher than for the general population. This may indicate that the kidney donation itself may be a factor in the evolution of the disease.

Or it could be a result of sampling bias. That is, kidney donors are more likely to have insurance and thus more likely to see a doctor who will diagnose the disease. The authors state,

“In our study, the increased prevalence of hypertension among Hispanic donors, as compared with the general population, may, in part, reflect underreporting of hypertension in this ethnic group, as compared with white respondents, in NHANES. We speculate that medical surveillance after kidney donation may mitigate barriers to the recognition of hypertension rather than differentially affect the risk of hypertension among Hispanic donors.”

 

Better advertising through behavioral research

When designing advertisements, creative directors often prepare mock ups of proposed designs to show to clients. Creative directors rely on their experience and training to prepare combinations of images and text that will hopefully engage readers. But they may not be able to articulate why their combination works or is better than another combination. Similarly, clients often accept or reject designs based on subjective and personal criteria. Is there a better way to judge the potential impact of an ad?

One method used by market researchers is eye tracking studies. This technique records a person’s eye position and movement when viewing visual media. You want the audience to be attracted to the image and headline which piques curiosity. Then they will be drawn into the text, and finally look at the company name, logo, tagline, and contact information.

The human eye and mind are remarkable devices. The human eye is not just a camera that records color and light. It also processes images before sending it to the brain. The eye is very slow at resolving an image. It takes about 1/20th of a second for the eye to generate an image to send to the brain. During that time, the eye must remain fixed on the point of interest, even if your body is in motion. This is called fixation. Then your eye moves quickly to the next point of interest using an action called saccade. During the saccade, your eye can’t create a good image, it is just a jittery blur. It doesn’t send this blurry image to the brain, it sends nothing. Yet you never notice that your entire visual life consists of a series of rapid still images with blackouts in between. In fact, that’s why movies and television (which consist of 24 or 30 still images a second) can fool you into thinking that there is constant motion.

Fixation and saccade. Image from Wikipedia

Eye tracking studies for advertising have been conducted for years, starting in the 1960s though as equipment improved and fell in price, the practice expanded. A new web service from 3M (Creativepro May 2010) called Visual Attention Service does not require actual consumer testing. Instead, this service uses a database of previous studies to predict what a person will look at and rate the effectiveness of the image and text in holding a person’s attention.

Visual Attention Service. Video from 3M

I’ve signed up as a user, but haven’t tried the service out myself yet. My guess is that the current tool is quite crude. But the concept makes sense and I can see that Google, Microsoft, and Yahoo! would be interested. Anything that increases the effectiveness of advertising is valuable and worth a lot to advertisers, publishers, and design firms.

****
While researching this blog entry, I came across a study by Think Eye tracking, a market research firm in Berkshire, UK, that reports the hilarious eye tracking results of a guy and what he looks at while attending a speed dating event. I think what actually happened is that he was so embarrassed and self-conscious about his appearance while wearing the eye tracking video camera headset that he had to avert his eyes. Yeah, that’s it.

And speaking of research into speed dating, previous research had indicated that women are pickier than men when selecting who to meet again. Two psychologists at Univ. of Penn. found that among 2,650 participants at HurryDate, the average woman was chosen by 49% of the men but the average man was only chosen by 34% of the women. However, this result may be biased because at most speed dating events, the man moves from table to table while the woman remains seated. A study by two psychologists at Northwestern in Psych. Sci. Sep 2009 shows that the gender selectivity difference disappears if women are the ones who rotate and men sit. The act of approaching someone increased self-confidence and reduced selectivity. Research bias can be very subtle.

[Update: Fixed a broken link to VAS YouTube video.]

 

Exploring kidney data (part 1) – transplant waiting list

The kidney transplant waiting list maintained by the UNOS gets longer every year as does the average waiting time for patients on the list. This is true even though the number of patients diagnosed with end-stage renal disease (ESRD) has declined slightly over the past few years. What is driving this? To examine this problem, I examined the data for annual changes in the number of patients with ESRD and the number on the waiting list.

About the data

The data for annual incidence (number of new cases of ESRD diagnosed in a year) and prevalence (total number of people with ESRD at the end of each year) were obtained from the USRDS 2009 annual report. The report contains a wealth of data on chronic kidney disease and ESRD. It has an entire chapter devoted to transplantation.

The data on the UNOS waiting list data was obtained from the UNOS. Some is available from OPTN annual reports or from the report builder web service. Others were generated specifically for me by UNOS. I want to thank Katarina Linden of UNOS for summarizing the SAS data used in this blog post. Any errors in analysis are mine alone.

A few notes regarding the UNOS data. First, the data being analyzed is for the kidney-only list. Patients are placed on lists based on what organ(s) they need. The UNOS maintains separate lists for each organ combination a patient needs, kidney only, kidney and pancreas, kidney and liver, etc. A single patient can be on more than one list. If the candidate receives a transplant, the transplant center is required to remove the patient from all the other lists as a duplicate entry.

Second, the counts are for registrations not candidates. In any year, about 5% of the kidney-only candidates (patients on the kidney-only waiting list) are registered at more than one transplant center. Most are people who have moved and are transferring their registration to a transplant center closer to their new address. But a few, most likely wealthy patients, are actually registered at multiple transplant centers in an effort to get an organ faster. The most famous example of this is Steve Jobs, who needed a liver transplant and had access to a corporate jet. But anybody who lives in a large city can benefit by getting on the list at a hospital in a more rural area, then traveling to that town and waiting for a donor after they reach the top of the list. Again, after the patient receives a transplant, all transplant centers are required to inform the UNOS that duplicate registrations for that patient should be removed from the list.

Third, there isn’t a direct correlation between the number of people on the UNOS waiting list and the number of people with ESRD (the prevalence rate). Once a patient receives a transplant, they are removed from the waiting list. However, they are not cured and so are still counted as having ESRD. Similarly, there is no direct correlation between the number of people added to the UNOS waiting list in a year and the number of people newly diagnosed with ESRD. That’s because a patient who enters the waiting list may have been diagnosed with ESRD years earlier. Also, they may enter the waiting list if the they previously received a kidney transplant and the organ fails.

Finally, there are two categories to the waiting list. Registrants are classified by the transplant center as either active and inactive. Active registrants are considered medically able to get a transplant immediately if an organ becomes available. Inactive registrants are currently unable to accept a transplant, but are considered good long-term candidates for a transplant. I will discuss this in more detail later.

The UNOS data is collected via a survey that each transplant center must complete for each registrant on their waiting list once a year to determine the registrant’s current status. As any of you who have dealt with survey data realize, cleaning survey response data from respondents (both the candidates and the administrators at the transplant center) who are not familiar with statistical analysis is one of the most difficult tasks in any research project and is a major source of nonsampling error.

Growth in prevalence of ESRD and in size of kidney transplant waiting list

In 2007, the latest year data is available, about 111,000 people in the U.S. were diagnosed with ESRD for an average incidence rate of 361 per million population. Figure 1 shows the incidence rate of ESRD has been rising dramatically over the past two decades, though it seems to have peaked. Blacks are more than three times likely to be diagnosed as whites. The ratio of prevalence by race is not as high, meaning that once diagnosed, whites tend to live longer with ESRD than blacks.

Figure 1. Incidence and prevalence rates for ESRD by race. Data from USRDS

In 2007, the prevalence of ESRD was about 1,700 per million population, representing over 527,000 people. Of these people, about 150,000 have a functioning transplanted kidney and 375,000 are on dialysis. (The remainder refuse treatment and will die within a few months.) However, as shown in Figure 2, there were only 78,300 registrants (and fewer candidates) on the UNOS transplant waiting list. This means only one-fifth of patients on dialysis were on the UNOS waiting list. As discussed in earlier blog posts (Dec 5 and Dec 18), the reasons people fail to get on the waiting list are complex.

Figure 2.Total active and inactive wait list. Data from UNOS

Figure 3 shows even though the number of transplants is growing (right), the incidence rate of ESRD is growing faster (left), so the wait times for a deceased donor kidney is getting longer (center) and the transplant rate is falling (left). For every 100 people newly diagnosed with ESRD in 2007, there were only four transplants.

Figure 3. Transplant trends. Data from USRDS

Growth in the UNOS waiting list

Figure 4 shows the number of new registrants being added to the active list each year rose from about 17,000 in 1995 to 25,000 in 2004 and has flattened out since then. Unfortunately, the number of transplants, from either deceased or live donors, has not kept pace. The seemingly good news is that the number of registrants removed from the active list without a transplant (which consists of people who decide they no longer want a transplant, are too sick for a transplant, or die) has not been growing. The categories that has been growing (and growing rapidly) are movements to and from the inactive list, with more patients going to the inactive list than coming from it. Notice the size of this churn represents a large proportion (more than one-fourth) of the active waiting list population.

Figure 4. Active wait list. Data from UNOS

Figure 5 shows the rapid growth of the inactive waiting list. First, notice the jump in the number of new inactive registrations starting in 2004. The most common reason for being initially placed on the inactive list is an incomplete evaluation by the transplant center. That is, the patient starts the evaluation process but is unable to complete it before the survey date (perhaps due to difficulty getting transportation to the transplant center). Other reasons for being on the inactive list are the presence of treatable comorbidities such as obesity, addiction (smoking, alcohol, or drugs), hypertension, or type 2 diabetes.

Next, notice the large number of registrants moving to the active list. Many of them are the new registrants who have completed their evaluations and are moved to the active waiting list, but also includes a large number of registrants who started as active and were on the inactive list for a period of time.

Most of the people on the inactive list are removed without a transplant, which makes sense, since they were not considered good transplant candidates. However, when coupled with the large flow of registrants from the active list (equal to about half of the total number of inactive registrants each year), this may also indicate that many transplant centers are moving active candidates to the inactive list rather than removing them entirely. Thus, the inactive list may contain many registrants who are too sick to receive a transplant and have little chance of recovery prior to death.

There are a large number of registrants in the Not coded category. This indicates that people on the inactive list are less likely to be in close contact with their transplant center and either could not be contacted or incorrectly completed the survey.

Figure 5. Inactive wait list, data from UNOS

****

In a future post, I will continue to explore the USRDS and UNOS data to revisit the issue of long wait times experienced by patients with type O blood.

[Update1: Added explanation that the counts of incidence and prevalence cannot be directly compared to the counts on the waiting list.]

[Update2: Corrected an error in proportion of patients with ESRD on the UNOS waiting list. Patients who have a functioning transplant should be excluded from the calculation.]

 

Is donating a kidney safe?

by George Taniwaki

The short answer is yes. The long answer is complex and interesting. (Well, it’s interesting if you are a statistics geek like me.)

First, some good news. Most of the available data indicates that live kidney donors lead long healthy lives. Studies show that they live longer than the general population. For instance, see Transpl. Oct 1997 and New Engl. J. Med. Jan 2009. This is not an unexpected result and does not mean that donating a kidney will lengthen your life. Instead, it is probably a result of the fact that kidney donors are screened for good health (called selection bias) and are healthier than the general population, and thus more likely to live longer.

Survival rate of kidney donors is similar to general population. Image from New Engl J Med

A more meaningful comparison would be to look at longevity of kidney donors compared to a stratified sample of the general population controlled for age, income, gender, geography, medical history, and access to health care (or health insurance). Such a study would be difficult to conduct. That’s because neither hospitals nor the United Network for Organ Sharing (UNOS) do a good job of tracking kidney donors after surgery. They do a better job of tracking recipients. So the data on the long-term outcomes of donors is sparse.

Donating a kidney does expose donors to several near-term risks that may shorten their lives. A study in J. Amer. Med. Assoc. Mar 2010 shows that in the 90 days after a donation, the mortality rate was 3.1 per 10,000 for donors compared to 0.4 per 10,000 for a control group. A good summary of these risks is provided by the Mayo Clinic and by the National Kidney Foundation. Actual risks may vary and donors should discuss them with the transplant surgeon. However, the risks are small, especially when compared to the great benefits that will be experienced by the recipients. Not all researchers are quite as sanguine. A note in Clinical J. Amer. Soc. Nephr. Jul 2006 cautions that more studies are needed.

For the long-term, the risk of premature death are low. The same JAMA study cited above shows the long-term survival is excellent. The risk of death was the same or lower than for the control group after five years (0.4% vs. 0.9%) and after 12 years (1.5% vs. 2.9%), respectively.

There is one risk that is correlated with kidney donation that is very odd and deserves additional investigation by epidemiologists. Specifically, it appears that kidney donors are more likely than the general population to develop end stage renal disease (ESRD). My friend, Ken Klima at Hebert Research, heard this surprising finding in a UWTV lecture entitled Understanding a Chronic Killer: Kidney Disease, Part 1 (additional kidney related videos are also available). The data is reported in a rather shocking manner by Wendell Fleet a professor of nephrology at the UWMC, which is where I am expecting to have my surgery. At 34:40 into the video, he says:

“If you donate one of your kidneys to a loved one, or in a fit of philanthropic zeal to a total stranger (audience laughs), you may wear out your kidney. We initially told people you only need one, ‘give it up and save someone’s life.’ So we followed those people and in a few of them the creatinine levels inch up. A few have required renal replacement therapy. They wore out their remaining kidney. On the positive side, you go directly to the top of the list for a transplant yourself if you give someone a kidney (audience laughs).”

Dr. Fleet may be referring to data from various studies, such as one reported in Transpl. Nov 2002 or Transpl. Proc. June 2008 (subscription required), that show that among patients undergoing living donor nephrectomies, about 0.35% developed ESRD compared to 0.25% for the general population. Although this is a difference of only 0.10%, it represents a huge increase of 40% (=0.10/0.25). Am I putting myself at risk for kidney disease by donating my kidney?

I don’t think so. I have a guess as to what’s really happening. Susceptibility to kidney disease is partially hereditary, as are other chronic conditions like diabetes and hypertension that are correlated with ESRD. Since historically a majority of kidney donors are family members of the recipient, they may have also inherited the genes that cause ESRD. A similar conclusion is stated in an editorial in the Nephr. Dialy. Transpl. May 2003. Again, a more complete analysis would compare rates of ESRD correcting for age, income, gender, geography, access to medical care, and medical history (especially a family history of ESRD).

For more information on becoming a kidney donor, see my Kidney donor guide.

[Update1: I added a link to a study that questions the low medical risks reported for live kidney donors.]

[Update2: I added a link to a new JAMA study.]

[Update3: An Aug 2010 blog post contains additional findings on the safety of kidney donation.]