In 2014, PAs began a transition to a 10-year certification maintenance cycle with specific continuing medical education (CME) requirements related to self-assessment (SA-CME) and to performance improvement (PI-CME).1 The rationale for these changes was threefold:
- Acknowledgment that traditional CME alone does not incorporate the true essence of continuous professional development that is integral to practicing as a competent healthcare provider
- All physician specialty boards (as represented by the American Board of Medical Specialties [ABMS]) implemented to a maintenance of certification process with similar requirements
- The previous certification maintenance model for PAs focused almost exclusively on medical knowledge and patient care, only two of the six competencies of the professional PA.1
PI-CME is a three-step process:
- Comparing some aspect of practice with national benchmarks, performance guidelines, or other established evidence-based metric or standard
- Using this comparison to develop and implement a plan for improvement in that area
- Evaluating the effect of the improvement effort by comparing the results of the original practice with the new results or outcomes.
Although PAs may meet the new CME requirements by completing PI-CME and SA-CME activities developed for physicians, there has been keen interest in the creation of PA-specific options to meet the requirements.
Chronic kidney disease (CKD) is a silent epidemic in the United States. The National Kidney Foundation (NKF) recently reported that Americans are “kidney clueless,” further stating that practitioners miss the diagnosis in more than 50% of high-risk patients.2,3 More than 60 million Americans (14% of the US population) have CKD and most do not know it.4 One in six Americans will develop stage 3 to 5 CKD (Figure 1) during their life-span.5,6 The cost to patients and their families is significant, yet the monetary cost to society is even higher. US taxpayers (under the auspices of the Medicare End Stage Renal Disease [ESRD] program) pay more for dialysis patients than they pay for the entire budget of the National Institutes of Health (NIH).7 The total in dollars of healthcare costs, lost productivity, and out-of-pocket expenses is staggering.
Patients often are referred to nephrology after kidney disease is advanced. However, early CKD identification and appropriate management by PAs can potentially slow disease progression and prevent patients from progressing to kidney failure requiring dialysis or transplant. Recognizing the need for PA-specific PI-CME projects, the American Academy of Nephrology PAs (AANPA) developed Kidneys in a Box (KIB). Because diabetes is the most common cause of kidney disease and failure, the stated goal of KIB was to improve management of diabetes earlier and consequently protect and preserve native kidney function.4
With input from the National Institutes of Health's National Kidney Disease Education Program (NIH/NKDEP), a list of six modifiable factors for patients with diabetes was developed:8
- statin use
- A1C measured within the last 6 months
- urine albumin-creatinine ratio (UACR) measured within the last year
- CKD stage
- a yellow caution over-the-counter (OTC) medication list to decrease iatrogenic kidney injury
- smoking cessation.
Participants were instructed to review 10 charts of patients with diabetes and determine if they did or did not meet national guidelines in these six areas. For example, did the PA put patients on statins, stage patients with 2013 CKD guidelines and/or order UACRs, share yellow caution medications or discuss smoking and kidney disease with patients? PA participants developed a practice-specific intervention for one of the six factors. The intervention was to be implemented for 3 months, after which the participant pulled 10 different charts of patients with diabetes. The PA participants determined if they had now met national guidelines on the six modifiable factors. The final evaluation sheet, along with a blinded preintervention/postintervention evaluation sheet, were returned to AANPA for PI-CME credit.
Two years after the introduction of the KIB program, data were collected and analyzed to ascertain if appropriate measures were implemented in nonnephrology practices. The long-term goal of KIB was to introduce permanent changes in practice.9
Six modifiable risk factors, those changes that have been shown to slow kidney disease in patients with diabetes, were selected after input from NKDEP. Information in the KIB program explained to participants how these six items could be incorporated in their practices. KIB PI-CME workshops were held in conjunction with multiple state PA meetings; the program also had a website where anyone could request the program. Selection of which factor and the method of change was left up to the participant who reported the project and outcomes to AANPA. In accordance with American Academy of PAs (AAPA)/NCCPA standards, each PA submitted blinded abstracted chart information describing present practice parameters of these six factors from 10 charts in the preintervention stage and 10 different charts in postintervention stages for a total of 20 patients. Upon completion of the KIB program, the preintervention/postintervention evaluation form, along with a demographic information sheet, was sent to AANPA to obtain CME credit. There was no outside review of the blinded charts and no attempt was made to link the chart data to the patient. Institutional review board approval was not required per the US Department of Health because data were submitted and stored in a deidentified form.10 The program launched in March 5, 2014, and by May 21, 2016, data were available on 302 PAs. Full preintervention/postintervention and demographic data were available for 213 PAs and these blinded data were analyzed by Timothy McCall, PhD, of the AAPA statistical department.
This study was designed to identify the effect of a PI-CME project on the management of patients with diabetes. The preintervention/postintervention reports along with two-tailed confirmatory tests of differences were analyzed for 213 participating PAs from 41 states and the United Kingdom. Final preintervention/postintervention forms were missing from 89 additional PAs and thus their demographic data were not used in the final analysis. No ascertainable difference was found in the 89 PAs who did not return the preintervention/postintervention data form and the study cohort. The largest cohorts of participants were from California, Florida, New York, and Pennsylvania, possibly due to higher number of workshops offered in those states. The average participant had been in practice for 10.7 years and saw 77 patients/week, of which 27 were patients with diabetes (Table 1). The study had a large cohort of specialty PAs, although 60% of the cohort identified as either internal medicine or family practice (Table 2).
Using paired samples, all six of the six modifiable factors were found to be more likely modified after KIB intervention, with five of the six showing statistical significance (Figure 2). A statistically significant increase was found in the amount of statins prescribed after the KIB intervention (mean=7.4, SD 2.2) versus before the intervention (mean=6.4, SD 2.1), t(147)=7.7, P<0.001. A statistically significant increase was found in the frequency of A1Cs and UACRs done after KIB intervention (mean=9, SD 1.6 and mean=7, SD 3.1, respectively) versus before the intervention ([mean=5.2, SD 3.4], t(147)=5.5, P<0.001 and [mean=3.6, SD 3.3], t(147)=9.8, P<0.001, respectively). A statistically significant amount of CKD staging was done after KIB intervention (mean=6, SD 3.2) versus before the intervention (mean=3.6, SD 3.3), t(147)=11.6, P<0.001. A statistically significant number of OTC lists were given after KIB intervention (mean=6.4, SD 3.3) versus before the intervention (mean=2.6, SD 3.5), t(147)=14.95, P<0.001. Although smoking intervention increased after the KIB intervention (mean=5.7, SD 3.1) versus before (mean=5.4, SD 3.1), t(147)=2.1, P=0.037, the number did not reach statistical significance (Figure 2).
CKD is becoming an epidemic in the United States. Many nonnephrology practitioners will see a patient with diabetes before the patient suffers irreplaceable loss of kidney function. This is the time when interventions can potentially slow disease progression. Although all clinicians hope that a patient is followed by a primary care or internal medicine provider who intervenes in CKD, frequently this is not the case.3 All too often, when it is clear that a patient's kidney function is declining (GFR <60 mL/min), many providers do not know which interventions are best to slow the progression. Some providers may feel that CKD is not in their realm of expertise and leave the interventions to specialists. The most common cause of kidney failure is diabetes, and KIB was developed to encourage improved screening of patients with diabetes for kidney disease and to encourage providers to implement methods to slow progression of CKD.4
KIB listed six modifiable risk factors for CKD in patients with diabetes. PAs in this study each reviewed 10 charts to see if patients met benchmarks as defined by KIB. If benchmarks were not met, the PA developed a practice-specific intervention to meet one of the six benchmarks. Although PAs in the study cohort were asked to select one of the six factors for intervention, many PAs selected more than one factor. The materials in the KIB toolbox let the PA use any of the patient and/PA education programs to develop an intervention, although the particular intervention was specific to the site, PA, and practice. This lets many different types of practices, specialties, and sites use the KIB program. The original expectation was that this program would be used exclusively by internal medicine or family practice PAs; however, 40% of the participants were in specialties.
Results show that KIB had a statistically significant effect on PAs by altering behavior for five of the six modifiable factors. Smoking cessation did not reach statistical significance in the study sample. An issue with wording of the question in the evaluation form was noted by multiple cohort participants and may explain the result. The evaluation sheet query stated: “Does the patient have a smoking history?” The purpose was to identify if a smoking history was in the medical record (previous/present/not applicable) and many participants believed the query was asking if the patient smoked now or previously. This may explain why there was no significant change on that particular risk factor. Also, smoking cessation may fall primarily on the shoulders of the patient. Yet, although the intervention did not reach statistical significance, several cardiology and vascular PAs initiated smoking cessation groups in their practices. All specialties reported referring patients to smoking cessation groups.
The five risk factors that showed statistical improvement (CKD staging, UACR, statin use, A1Cs, and OTC yellow medication lists) provide an immediate opportunity for intervention. These are changes that can be accomplished while the patient is in the clinic or office or via referral.
Many providers know to follow patients with diabetes with routine hemoglobin A1C and cholesterol levels. The UACR will show albumin in the urine before the drop in GFR, making urine screening vital.8 However, although patients with diabetes are known to be at higher risk for developing kidney disease, fewer than 40% of these patients are tested for protein or albumin in their urine (Figure 3). Even more shocking, only 42% of patients diagnosed with CKD had urine albumin/protein checked in the previous year despite national guidelines requiring UACR testing annually for patients with diabetes (Figure 4).
With the move to ICD-10, staging kidney disease in a patient's medical record prompts clinicians to obtain a urinalysis for albumin, and many of the PAs noted that their electronic health record was adapted by their information technology departments at the same time as they were working on KIB. Thus, the modifiable risks with the greatest improvement in clinical practice included obtaining an UACR and documenting a stage for CKD.
The most popular intervention was providing a list of yellow caution nephrotoxic OTC medications to patients; many of the cohort PAs developed handouts specific to their populations, including translation into multiple languages. KIB developed and supplied OTC charts that included common OTC medications, herbals, Navajo Indian remedies, and probiotics. These charts, which were developed with input from multiple sources, listed common usages and precautions for patients with CKD by stage of kidney disease. A significant number of PAs noted they did not realize so many OTC medications could affect their patients with CKD.
We found that the PA community embraced the chance to make substantial changes to management of patients with diabetes. Comments collected on the evaluation forms were overwhelmingly positive. Specialty PAs were pleased to make changes in their practices in regards to CKD. A large number of PAs were fascinated by the information and stunned to learn the depth and breadth of the CKD crisis. The outreach outside of the goal audience of primary care and internal medicine was unexpected and yet judging by comments on evaluations, seemed to make the largest difference. Outreach to specialty PAs and specialty practices should continue to be a goal of the nephrology community as many of the specialty PAs noted their patients did not have a designated primary care provider.
Limitations of this study include the self-reporting nature of PI-CME, the small sample size, the short follow-up time, and the implementation of ICD-10 during the intervention. Also, because the intervention was short term (3 months), whether permanent changes were made in the behavior of the PAs remains a question. The ICD-10 codes were implemented in the 6 months before the final collection of data from KIB. However, the minimum time required for a PI-CME project is 3 months and thus the overlap with the change to ICD-10 may be minimal. Data can be skewed if the 89 PA participants who did not turn in the preintervention/postintervention form were significantly different from the cohort PAs, although their demographics looked similar to the cohorts. Data can further be skewed if only those who succeeded in their intervention returned the evaluation forms. However, unless the form is returned, no CME credit is obtained and the program notes that a negative result earns credit just as does a positive result. Thus, we feel this is unlikely to change results. Strengths include the diverse community in terms of geographic areas, types of practices, and years in practice, along with statistically significant changes that occurred in practices.
Before this PI-CME, interventions in quality projects have been reported in cardiology, podiatry, dermatology, pain management, osteoporosis, and treatment of depression.11-17 Although more than 26 million Americans have CKD and the expenses associated with this population are astronomical, most previous interventions in CKD have been addressed to primary care and with limited success. By using the outreach materials already developed by NKDEP and building on the need for PAs to obtain PI-CME credits, KIB was able to reach a large number of PAs nationally. KIB demonstrated statistically significant behavioral changes in five of the six modifiable risk factors. This was true for all types of practices—specialty, primary care, and internal medicine. CKD takes years to develop but the butterfly effect leads one to believe that simple interventions may truly slow CKD progression.
1. Cohen PM. Building the right skills for the healthcare workforce of the future. JAAPA
2. National Kidney Foundation. Americans are kidney clueless. https://www.kidney.org/news/americans-are-kidney-clueless
. Accessed July 11, 2016.
3. Szczech LA, Stewart RC, Su HL, et al. Primary care detection of chronic kidney disease in adults with type-2 diabetes: the ADD-CKD Study (awareness, detection and drug therapy in type 2 diabetes and chronic kidney disease). PLoS One
4. US Renal Data System. 2015 annual data report: epidemiology of kidney disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2015.
5. Grams ME, Chow EK, Segev DL, Coresh J. Lifetime incidence of CKD stages 3–5 in the United States. Am J Kidney Dis
6. National Kidney Foundation. CKD facts. https://www.kidney.org/news/newsroom/factsheets/FastFacts#Ref
. Accessed July 11, 2016.
7. Cho S. ASN president urges kidney disease innovation. Nephrology News and Issues
, April 9, 2014. www.newswise.com/articles/asn-president-calls-for-kidney-disease-innovation-in-congressional-testimony
. Accessed July 11, 2016.
8. National Kidney Foundation. KDOQI clinical practice guideline for diabetes and CKD: 2012 update. Am J Kidney Dis
9. Silver SA, McQuillan R, Harel Z, et al. How to sustain change and support continuous quality improvement. Clin J Am Soc Nephrol
10. US Department of Health and Human Services. Code of Federal Regulations. Title 45. Public welfare. Part 46. Protection of human subjects. Rev 2009. www.hhs.gov/ohrp/policy/ohrpregulations.pdf
. Accessed July 11, 2016.
11. Gist DL, Bhushan R, Hamarstrom E, et al. Impact of a performance improvement CME activity on the care and treatment of patients with psoriasis. J Am Acad Dermatol
12. Thase ME, Stowell SA, Berry CA, et al. A performance improvement initiative for enhancing the care of patients with depression. J Psychiatr Pract
13. Joyner J, Moore MA, Simmons DR, et al. Impact of performance improvement continuing medical education on cardiometabolic risk factor control: the COSEHC initiative. J Contin Educ Health Prof
14. Szpunar SM, Minnick SE, Dako I, Saravolatz LD 2nd. Improving foot examinations in patients with diabetes: a performance improvement continuing medical education (PI-CME) project. Diabetes Educ
15. Greenspan SL, Bilezikian JP, Watts NB, et al. A clinician performance initiative to improve quality of care for patients with osteoporosis. J Womens Health (Larchmt)
16. Mullikin EA, Ales MW, Cho J, et al. Sharing collaborative designs of tobacco cessation performance improvement CME projects. J Contin Educ Health Prof
. 2011;31 (suppl 1):S37–S49.
17. Fine PG, Bradshaw DH, Cohen MJ, et al. Evaluation of the performance improvement CME paradigm for pain management in the long-term care setting. Pain Med