NephJC: Educational programs improve the preparation for dialysis and survival of patients with chronic kidney disease.
The accompanying editorial is free. Get the PDF here.
Live TweetChat: April 29th at 9PM Eastern
This is an observational study regarding the NKF-funded KEEP program. Briefly, KEEP (Kidney Early Evaluation Program) is a national, community based screening and education program launched in August 2000. KEEP is a screening program that identifies people at high risk of kidney disease, and encourages them to follow-up with kidney focused preventative care. High risk is defined as:
- Self-reported history of diabetes or hypertension
- First degree relative with hypertension, diabetes or kidney disease
In addition to the personal and family history, participants have a serum creatinine and urine albumin:creatinine ratio checked. They then meet a physician and review their risk factors and discuss possible interventions.
For each participant, the final step is a letter describing test the results, the individuals presense or risk of future kidney disease and a personalized clinical action plan for treating CKD. This is sent to the participant and her doctor.
Patients who participated in KEEP and subsequently developed ESRD made up the intervention arm (N = 595). For the comparator the investigators used the USRDS database to identify incident ESRD patients who started dialysis during the same time period and who lived in the same zip-code. Propensity score matching was used (in a ratio of 5 controls for each KEEP participant) to identify 2975 non-KEEP control patients.
The baseline demographics (table 1) do show that the initial USRDS sample of 290,252 who started dialysis during that period were different in many ways from the KEEP participants; after the propensity score matching though, this sample was whittled down to a fairly comparable cohort with no significant imbalances. Amongst the results (Table 2), the KEEPers were:
- More likely to have seen a nephrologist (76% vs 69%)
- More likely to have been placed on the transplant waiting list before ESRD (24% vs 17%)
- More likely to have undergone transplantation (10% vs 6%)
- More likely to use peritoneal dialysis (10% vs 6%)
There were also some other interesting trends that did not reach statistical significance, KEEPers were more likely to:
- Use an arteriovenous fistula or graft (23% vs 20%)
- Have a mature or maturing AVF/AVG at first dialysis (44% vs 40%)
The most important finding was that mortality was lower (HR 0.80, 95% CI 0.68 — 0.94), though this was attenuated after adjustment (table 3).
The results of this study are extremely interesting for the nephrology community as we have struggled with improving the outcomes in our patients. As the accompanying editorial remarks,
During the chat we embedded some of the key tables in to a tweet. Here are the tweets:
As we are trying to build a style, we'd like to know, do you think this is a good practice or not?
The First #NephJC: Wrap Up
The Live chat
After all the build-up, the first nephrology journal club, went live on April 29th at 9 pm EDT (in case you were sleeping under a rock somewhere!). For the first 15 seconds, the two of us were the only ones signed in and we had a few anxious moments:
But the nephrology tweeps rose to the occasion. Ultimately, just over 20 participants joined the chat, and there were more watching. The tweets came fast and there were multiple simultaneous conversations. There were many disagreements (Which model was most appropriate? is a trial ethical?) but the points were made with reason and in a polite, respectful tone (what can you expect with a Canadian involved, eh?). Amongst many of the highlights was the active participation of @AfternoonNapper, an e-patient of some regard and future KidneyWeek speaker, injecting a unique insight into the discussion.
The hour passed quickly, and at the end, more than 200 tweets had been sent. Even after the chat was over, reactions ‘KEEP’ coming and many more tweeps promise to join in next time.
We had three broad topics for discussion.
The first topic regarded the internal validity of the study. There was near unanimous agreement that selection of the control group is difficult, and that voluntary participation in KEEP may be a marker for healthier, more engaged or more educated individuals (who would perhaps make better choices and decisions). The KEEP investigators are to be commended for doing a meticulous and painstaking propensity score matching analysis to compensate. But the crowd was not convinced that this statistical slight of hand eliminated this fundamental bias.
The second discussion point about the interpretation of the various models was quite a bit more contentious, and there was no consensus (see below for more details and the authors’ take).
Lastly, only a few hardcore EBM types thought there was enough equipoise to even consider doing a randomized controlled trial to test the hypothesis of whether screening and education for CKD makes a real difference, unlike the editorialists suggestion. Overall, everyone expressed their appreciation for the KEEP investigators, and the National Kidney Foundation for funding it.
Author Q and A
We sent a follow-up e-mail to the authors with some of the questions and issues raised in the Journal Club. Here is an edited transcript with the lead author, Dr Manjula Kurella Tamura (Dr. MKT), from Stanford.
NephJC: In Table 3 there are a number of models, and there was a disagreement on which is most valid. One opinion was summed up in this tweet:
and the other position in this tweet:
Dr MKT: We presented several models here precisely because of the issues articulated during your discussion. I am happy to let readers decide which model (if any) is most valid. Let me present one additional interpretation that may not have been raised in your discussion: The association between KEEP participation and reduced mortality was attenuated after adjusting for ESRD preparation, perhaps suggesting that improved ESRD preparation mediated some of the lower mortality effect.
NephJC: There were some questions on the unconventional end-point considering this was a pre-dialysis intervention. Nephrologists are used to seeing a doubling of creatinine or dialysis in this setting. Very few papers look at dialysis end-points in pre-dialysis interventions. There was a bit of curiosity about how many of the KEEP participants progressed to dialysis as opposed to, say, a cohort from NHANES or CRIC, which have no intervention.
Dr MKT: I agree that survival bias is an important concern that we could not address in this study. Ultimately, it will be important to perform the comparisons suggested with NHANES and/or CRIC, but when our study was conducted, NHANES data from 2005-2010 had not yet been linked with the National Death Index or USRDS.
NephJC: The last question was a debate of whether there needed to be an RCT to fully answer this question or are we morally prevented from screening patients and then not informing half of them of the results. Seems ethically questionable?
Dr MKT: Maybe a more defensible design is to randomize patients to screening vs. not? Those randomized to screening would find out the results, those not randomized to screening would have no results. Clinicians could do whatever testing they felt was indicated during the course of clinical care in either arm. Equipoise could be defended based on ACP’s recent position statement against screening in adults without CKD risk factors.