Primary outcomes
Provider knowledge . Previous analysis already compared providers’ pre-test level of transplant knowledge, participating dialysis providers had significantly increased transplant knowledge following the in-person or webinar trainings. The 13 transplant knowledge questions included 4 true/false questions (e.g., ‘Living donors have a risk of blood pressure after donating’), 1 open-ended question (i.e., ‘Dialysis does what percent of the work of one functioning kidney?’), and 8 multiple choice questions (e.g., On average, how many years is a kidney transplant from a living donor expected to last?’).Knowledge score was tested using paired t-tests or McNemar tests.
Provider readiness and confidence . Previous analysis showed that immediately after ET training, 88% of dialysis providers plan to use Explore Transplant after the Trainings while 33% actually did. Pre- and post-training, dialysis providers will be asked how ready they are to educate 5 transplant candidates about deceased and living donation using the following scale:
‘I am not ready to educate 5 transplant candidates (Precontemplation)’, ‘I am considering whether to educate 5 transplant candidates (Contemplation)’, ‘I am preparing to educate 5 transplant candidates (Preparation)’, ‘I am in the process of educating 5 transplant candidates (Action)’, or ‘I already have educated 5 transplant candidates (Maintenance)’.
They were also asked to rank 5 statements about their ability to conduct transplant education (e.g., “I am sufficiently knowledgeable about transplant that I could answer more patients’ questions.”) using a 4-point Likert scale ranging from 1 (Strongly Disagree) to 4 (Strongly Agree). Chi-squared analyses were used to assess whether providers increased in their readiness to educate and their agreement that they are well-prepared to educate.
Difference-In-Difference analysis of WL rates pre-post ET. Previously in the paper draft, Devin found evidence that Explore Transplant increased dialysis centers’ living donor transplant rates: Controlling for covariates, dialysis centers where the provider educated 5 or more patients using ET in the subsequent four months were 1.33 times more likely to have at least one patient get a LDKT in 2010 and 2011 than dialysis centers where the provider educated less than 5 patients. There was also evidence of geographic disparities in LDKT rates: dialysis centers located in urbanized areas (> 50,000 pop.) were 1.65 times more likely to have at least one patient get a LDKT in 2010 and 2011 than dialysis centers located in urban clusters or rural areas (< 50,000 pop.). However,no evidence of within-center effects were found; there was not a significant difference in centers’ likelihood of having a patient receive LDKT between 2010 and 2011, indicating that while there is a difference in the likelihood of having a patient receive an LDKT if the provider educates > 5 patients with ET, this likelihood does not increase or decrease over time.
To estimate a causal effect of ET on WL and LDKT rates, we’ll compare CHANGES in WL rates and LDKT Pre and post ET between dialysis centers that attended ET vs those that did not send representatives to attend via Difference-In-Difference analysis.
variables we need:
A dummy variable to identify the ET group vs. non ET group.
A time variable indicating pre vsw. post using 0 and 1.
Create an interaction between time (2010-2011) and treated, and call this interaction “did”.
Combine pre and post rates into 1 singe variable and use time variable to indicate pre (0) vs. post (1).
Estimating the DID estimator: m.did = lm(Rates_12m ~ ET + time + did, data = mydata)
summary(m.did)
My questions:
- Where is the dataset that inlcudes centers that didnot attend ET and,
- Does the USRDS dataset include those? If so, we need to first merge them together
ET live vs. webnar effectiveness in promoting provider knowldege, motivation, and readiness etc.. At pre-test, we will collect data on provider and dialysis center characteristics including: gender, age, race, job title, length of time working with dialysis patients, current transplant education practices, number of dialysis centers covered, the presence or absence of barriers to conducting transplant education in their centers, and total patient caseload. We will also track which type of Explore Transplant training the provider attended (e.g., in-person or webinar).
We hypothesize that dialysis providers who participate in the in-person trainings will be more knowledgeable, ready to educate, and motivated to educate 5 patients within the next six months compared to the providers completing the webinar. To test this hypothesis, we will conduct chi-square tests of independence and one-way ANOVA tests to identify whether demographic, employment characteristics, characteristics of the transplant center, or type of Explore Transplant training received were associated with greater post-training transplant knowledge, readiness and motivation to educate 5 patients about transplant. We will conduct linear regression to predict level of transplant knowledge by demographic variables, level of transplant knowledge, perceived benefits and barriers to education within dialysis centers, and type of Explore Transplant training received. We will conduct logistic regression to predict whether providers plan to educate 5 patients about transplant in the next six month vs. not by demographic variables, level of transplant knowledge, perceived benefits and barriers to education within dialysis centers, and type of Explore Transplant training received. We will examine all of the possible predictors through univariate analyses first and enter only the predictors identified as significant (p<.05) into the final regression models.
What we need:
- A variable to code for ET live vs. Webinar.