Significance Statement
Although most patients with CKD are managed in the primary care setting, the evidence base for CKD care in general practice is scant, and it is not clear whether specific CKD management programs can alter outcomes in primary care. The authors conducted a cluster randomized, controlled trial comparing 23 primary care practices offering usual care with 23 primary care practices in which a nurse practitioner helped those practices interpret patient data files and implement guideline-based CKD interventions. They found that the intervention did not affect rate of eGFR decline, but it did lead to significant improvements in processes, quality of care, and the number of patients achieving BP targets. This approach may improve processes of care, potentially reducing the burden of cardiovascular disease in patients with CKD.
Keywords: chronic kidney disease, primary care, nurse intervention, clinical trial
Visual Abstract
Abstract
Background
Most patients with CKD are managed in the community. Whether nurse-led CKD management programs improve outcomes in patients with CKD in primary care is unclear.
Methods
To assess the effect of such a program on the rate of renal function decline in patients with CKD (stages 3–5) in primary care in the United Kingdom, we conducted a cluster randomized trial, the Primary-Secondary Care Partnership to Improve Outcomes in Chronic Kidney Disease study. A software program designed for the study created a data file of patients with CKD in participating practices. In 23 intervention practices (11,651 patients), a CKD nurse practitioner worked with nominated practice leads to interpret the data file and implement guideline-based patient-level CKD management interventions. The 23 control practices (11,706 patients) received a data file but otherwise, continued usual CKD care. The primary outcome was defined at the cluster (practice) level as the change from baseline of the mean eGFR of the patients with CKD at 6-month intervals up to 42 months. Secondary outcomes included numbers of patients coded for CKD, mean BP, numbers of patients achieving National Institute for Health and Care Excellence BP targets for CKD, and proteinuria measurement.
Results
After 42 months, eGFR did not differ significantly between control and intervention groups. CKD- and proteinuria-related coding improved significantly along with the number of patients achieving BP targets in the intervention group versus usual care.
Conclusions
CKD management programs in primary care may not slow progression of CKD, but they may significantly improve processes of care and potentially decrease the cardiovascular disease burden in CKD and related costs.
CKD is a significant worldwide public health problem. Depending on the ascertainment methodology used, the reported United Kingdom prevalence of CKD stages 3–5 in adults ranges from 5.7% to 8.5%.1,2 Individuals with CKD have increased risks of all-cause mortality, cardiovascular (CV) mortality and morbidity, and ESRD requiring dialysis or transplantation.3 Many patients with CKD have other long-term conditions, such as hypertension and diabetes mellitus, but the major driver of their increased CV risk is the prevailing kidney disease itself rather than associated comorbidities.4
To assist in management of CKD, in the last 10 years, serum creatinine–derived formulaic eGFR have been widely introduced to allow for accurate classification of patients by CKD stage5 and to support guideline-based management of patients with CKD by nonspecialists.6,7 In the United Kingdom, the introduction of a Renal National Service Framework8 mandated seamless CKD treatment pathways supported by financial incentives for primary care9 to achieve key targets in CKD management. These changes have been somewhat controversial due to the attendant increase in workloads for health care workers; doubts about the accuracy and relevance of eGFR, especially in the elderly population; and a lack of underpinning evidence of efficacy, particularly in primary care.10–13 Many clinicians in primary care remain uncomfortable with the concept of CKD and find this area of patient management challenging, difficult, and of debatable benefit.10 Consequently, engagement of primary health care teams in CKD care has been suboptimal.14
There are fewer randomized, controlled trials in nephrology than in other medical specialty areas.15 The evidence base for the management of individuals with pre-ESRD CKD is derived largely from young (<60 years old), heavily proteinuric, well clinically phenotyped patients with more advanced CKD in secondary care.16,17 However, most patients with CKD receive only primary care without referral to a secondary care nephrologist. These patients largely have CKD stage 3 and are older (>70 years old),2 with absent or minimal proteinuria and no clearly defined cause for their CKD. Clinical guidelines for CKD care by nonspecialists, therefore, rely on extrapolation from studies in patient groups where the pathophysiology may be substantially different. Whether nascent primary care CKD management pathways can effectively deliver reductions in progressive CKD and/or CV morbidity, thus reducing costs in secondary care, has not been well studied in large trials.
Primary care in the United Kingdom is extensively supported by practice-based clinical information technology (IT) systems, with the potential to provide a rich data source to support CKD management if the appropriate queries can extract CKD-relevant data and present it to health care workers in a usable format. The Primary-Secondary Care Partnership to Prevent Adverse Outcomes in Chronic Kidney Disease (PSP-CKD) study used primary care–based informatics to support a clinical trial designed to examine the ability of a nurse-led intervention on the basis of published CKD guidelines, delivered in primary care but supported by secondary care nephrologists, to mitigate the rate of decline in renal function in the typical patients with CKD encountered in primary care.
Methods
Study Design and Stakeholder Engagement
The PSP-CKD study (ClinicalTrials.gov identifier: NCT01688141) is a cluster randomized, controlled trial of a nurse specialist–assisted CKD management program in primary care (Supplemental Material has the full trial protocol). The PSP-CKD study is part of a portfolio of applied health research studies into chronic conditions supported by the National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care Leicestershire Northamptonshire and Rutland (CLAHRC LNR), and it has a pragmatic design to facilitate smooth integration into the varied clinical models used to deliver care of long-term conditions in different United Kingdom general practices. Study design and feasibility were guided by a study team, including clinicians (doctors and nurses) from both primary and secondary care, clinical managers, a public health expert, a statistician, and patient representatives. Before study approval, an external peer review of the PSP-CKD study protocol was conducted by the multidisciplinary CLAHRC LNR Scientific Committee, which also included patient representatives. The study team continued to meet regularly during the study.
Participants
All 69 primary care general practices in the English county of Northamptonshire are members of a single Clinical Commissioning Group (CCG) organized into locality subgroups, and around 400 general practitioners provide primary care for the population of approximately 650,000. To generate interest in the study, an engagement program of the PSP-CKD study was presented to the CCG Board, locality groups, and individual practices, delivered by the Chief Investigator and study team. All practices in Northamptonshire were then invited to participate. Clustering was at the level of the individual general practice, and patient consent was not sought. Individual general practices were randomized 1:1 to either receive a structured nurse practitioner–led CKD management program with secondary care nephrology support (the intervention group) or continue with usual primary care, general practitioner–led CKD management (the control group).
Study Structure
Informatics
To accurately identify all patients with CKD in each practice, a web-based CKD management and audit software tool, Improving Patient Care and Awareness of Kidney Disease Progression Together2 (IMPAKT; www.impakt.org.uk), was designed and developed to support the study, and it was configured to be compatible with all prevailing practice IT systems. IMPAKT uses Morbidity Information Query and Export Syntax search methodology to execute queries and extract Read coded CKD-relevant data from practice IT systems.
IMPAKT retrospectively searched the electronic health records of all registered individuals age >17 years old in each enrolled practice in both control and intervention groups to identify all patients with one or more eGFR<60 ml/min per 1.73 m2 in the previous 5 years. All prevalent patients receiving maintenance dialysis or with a renal transplant were excluded. For all identified patients with CKD, IMPAKT extracted additional demographic data (including age, ethnicity, past medical history, smoking history, body mass index, and CKD-relevant medication history) and biomedical data (including BP records, blood biochemistry and hematology results, and urine results relating to proteinuria). IMPAKT further analyzed the CKD data to determine the accuracy of existing CKD coding in the practice record and identify any uncoded/unrecognized patients with CKD. IMPAKT audited management of individual patients with CKD against National Institute for Health and Care Excellence (NICE) guideline standards for BP control, proteinuria management, and repeat measurement of eGFR within recommended timeframes. In addition, individual patients were stratified according to the number of the following risk factors at that they possessed: no recent eGFR, latest BP >150/90 mm Hg, proteinuria with urine protein:creatinine ratio (PCR) >50 mg/mmol, total cholesterol >6 mmol/L, diabetic, eGFR<45 ml/min per 1.73 m2, evidence of linear eGFR decline over time, hemoglobin <105 g/L, black/Asian ethnicity, current smoker, a coded history of CV disease, or a coded history of bladder outflow obstruction.
After interrogation of practice IT systems, a Microsoft Excel–based IMPAKT CKD data file was downloaded onto the practice network containing data relating to all patients with CKD in the practice with patient identifiers to allow for active management of this group of patients. An identical but pseudonymized data file was exported securely to the Leicester Clinical Trials Unit. IMPAKT was run at entry of each practice in the trial (T0) and then again at 6-month intervals until the end of the study.
Intervention
The research team downloaded an IMPAKT CKD file onto the practice network and orientated the practice clinical staff to the location of the file, but thereafter, in control practices, they provided no further input into the interpretation of the file data or the management of any patients with CKD identified by IMPAKT. In contrast, intervention practices nominated a CKD clinical lead, either a doctor or a nurse, and received the support of an experienced CKD nurse practitioner who worked in the practice with the nominated lead to interpret the IMPAKT CKD data file and then, implement patient-level CKD management interventions on the basis of the data supported by local secondary care nephrologists where necessary. CKD nurses were each allocated a portfolio of practices and then encouraged each allocated practice to assess their barriers to the application of best practice in CKD management. The nurse-led intervention was designed to be flexible according to the wishes of each practice. Thus, the intervention included any or all of correct clinical coding of patients with CKD, exploring ways of integrating better CKD care into day to day practice, assistance in implementing CKD guidelines, patient recall for testing, assistance in implementing BP and other CV risk factor management guidance, helping with medicines management, delivering dedicated primary care CKD clinics, and targeted management of those patients with CKD with highest number of risk factors (see above). Typically on visits to intervention practices, nurse practitioners would access clinical systems to implement an agreed plan using the IMPAKT CKD data file to discuss patient-level data with practice CKD leads and other key staff, ascertain patients with CKD to update practice disease registers, identify and recall patients with CKD due for proteinuria testing and/or repeat eGFR measurements, intensify BP management for those patients not at target, and ensure that modifiable CV risk factors were appropriately managed. In some practices, intervention nurse practitioners were permitted to select those patients at highest risk for CKD progression and/or CV disease and review them individually in an in-practice CKD clinic. Advice provided to the intervention group of practices was in line with NICE CKD guidelines.8
IMPAKT also has an in-built audit function configured to provide contemporaneous practice-level CKD management performance data in PowerPoint file format. To maintain practice engagement with the PSP-CKD study, intervention nurse practitioners presented rolling CKD audits at practice clinical meetings, and a member of the practice clinical staff was encouraged to adopt the CKD audit for continuing professional development.
In the intervention group, CKD nurse practitioners made phone contact with their allocated practices at least once weekly and visited at least twice monthly in person. In control practices, the only contact with the research team was a 6-month visit to run IMPAKT on the practice IT system. Nurses maintained a reflective diary record of all practice contacts. The intervention period lasted for 42 months, and data downloads were performed on eight occasions in each practice (T=0, 6, 12, 18, 24, 30, 36, and 42 months).
Outcomes
The primary outcome was defined at the cluster (practice) level as the change from baseline of the mean eGFR of the patients with CKD at 6-month intervals up to 42 months. Secondary outcomes were numbers of patients coded for CKD, mean BP, and numbers of patients achieving NICE BP targets for CKD.
Sample Size
The study was powered to detect a difference in mean eGFR of patients with CKD between practices of 3 ml/min per 1.73 m2 after 3.5 years of follow-up. To have 80% power to find such a difference significant at the 5% level (two-tailed test), assuming a within-subject SD of 12 and an intracluster correlation coefficient conservatively estimated to be 0.1 on the basis of the work by Campbell et al.18 and local primary care audit data (J. Mason, personal communication), 10,394 patients per group were required or a total of 25,985 after allowing for 20% attrition.
Human Subjects Protection
The study was sponsored by the University Hospitals of Leicester National Health Service (NHS) Trust (UHL), and it was approved by the Leicestershire, Northamptonshire, and Rutland Research Ethics Committee 1 (reference number 08/H0406/117) and the UHL research governance department.
Randomization
A statistician working at Leicester CTU but independent of the PSP-CKD study team generated a randomization list. After practices gave written consent to participate, a member of the research team performed an IMPAKT download at the practice and then, contacted Leicester CTU by phone with a practice code. The practice was then randomized by the CTU statistician using the randomization list. Individual general practices could not be blinded to their allocation.
Statistical Methods
Demographic variables are summarized by arm and overall. The difference in mean eGFR change from baseline between groups was analyzed at the practice level using linear multilevel mixed effects models, with all follow-up values included in the model and a random practice effect.
This analysis was conducted on a complete patient basis, with no imputation for mean eGFR values that could not be calculated for a given practice and follow-up period. Initial models included only a dichotomous intervention covariate. Adjusted change models were also fitted with an additional baseline mean eGFR covariate. A secondary analysis was carried out for the primary outcome, fitting linear regression models to analyze the change in mean eGFR from baseline to each follow-up time by intervention group separately. Where data allowed, the primary analysis was repeated for subgroups with CKD stages 3a and 3b. The primary analysis was repeated for eGFR in the subgroup of patients who had a precalculated negative eGFR gradient at baseline. The primary and secondary analyses were repeated for continuous secondary outcomes. For categorical variables, primary and secondary analyses were performed using Poisson multilevel mixed effects models and Poisson regression models, respectively, with the additional inclusion of an offset term for the total number of patients in a practice at each time point. Read code efficiency for CKD diagnosis and proteinuria assessment was explored by comparing numbers of patients diagnosed via Read code with those diagnosed via eGFR for CKD and ACR/PCR measurement or urine dipstick result for proteinuria. All analyses were conducted in SAS 9.4 and/or Stata 14.
Estimation of Intervention Costs
Costs were estimated from the perspective of the NHS on the basis of intervention costs of one specialist nurse practitioner working full-time across ten practices for 1 year as in the trial. Associated costs, such as additional appointments, or savings, such as from reduced CV events, were excluded. The IMPAKT software itself was not costed, and any transport costs for the nurse traveling between practices were excluded. Costs are shown in 2015/2016 GBP.
Data Sharing
All deidentified individual participant data collected during the trial together with the study protocol, the statistical analysis plan, and analytic code will be made available to any researchers who provide a methodologically sound proposal to achieve the aims in the approved proposal. Such proposals should be directed to njb18@le.ac.uk. The data will be made indefinitely available immediately after publication. To gain access, data requestors will need to sign a data access agreement.
Results
Forty-nine individual general practices with 353,256 registered patients >17 years of age were recruited and randomized. No practices withdrew, although three pairs of practices merged during the study, leaving 46 practices at the end of the trial, 23 in each arm. The mergers took place between practices in the same arm of the study, and thus, there was no crossover between groups. There was no difference in median deprivation score between groups. All practices completed follow-up to 36 months, and 30 practices completed follow-up to 42 months.
At baseline in the 49 practices, 31,056 individuals had either a single eGFR value of <60 ml/min per 1.73 m2 or a coded diagnosis of CKD, and of these, 23,357 had two or more eGFR values <60 ml/min per 1.73 m2 >3 months apart. Only this latter group was included in the study. The baseline demographics of the PSP-CKD study are shown in Table 1. Control and intervention groups were generally well matched. At the cluster level, similar numbers of individuals were split across the control and intervention arms, and individual cluster sizes were similar. Overall, age, body mass index, and proteinuria were similar in each arm. Diagnoses of diabetes mellitus were higher in the control group, but hypertension and CV disease were similar across the groups.
Table 1.
Summary of Primary-Secondary Care Partnership to Prevent Adverse Outcomes in Chronic Kidney Disease study baseline characteristics of control and intervention groups
| Characteristic | Control | Intervention | Total |
|---|---|---|---|
| Patients with CKD stage 3–5, n (%) | 11,706 (50.1) | 11,651 (49.9) | 23,357 |
| No. of practices, n (%) | 23 (50) | 23 (50) | 46 |
| Cluster size, median (IQR) | 548 (396–715) | 507 (269–655) | 521 (325–662) |
| Women, % | 7285 (62.2) | 7234 (62.1) | 14,519 (62.2) |
| Age, yr, mean (SD) | 75.4 (11.3) | 75.1 (11.4) | 75.3 (11.3) |
| BMI, kg/m2, mean (SD) | 28.3 (5.7) | 28.3 (5.7) | 28.3 (5.7) |
| Ethnicity, n (%) | |||
| White | 7894 (67.4) | 5601 (48.1) | 13,495 (57.8) |
| South Asian | 114 (1.0) | 98 (0.8) | 212 (0.9) |
| Black | 106 (1.0) | 128 (1.1) | 234 (1.0) |
| Other | 70 (0.6) | 53 (0.5) | 123 (0.5) |
| Missing data | 3522 (30.1) | 5771 (49.5) | 9293 (39.8) |
| Mean average eGFR, ml/min per 1.73 m2 per cluster (95% CI) | 53.6 (52.7 to 54.5) | 54.6 (53.6 to 55.5) | |
| CKD stage (MDRD), n (%) | |||
| 3a | 8934 (76.3) | 8979 (77.1) | 17,913 (76.7) |
| 3b | 2226 (19.0) | 2076 (17.8) | 4302 (18.4) |
| 4 | 466 (4.0) | 523 (4.5) | 989 (4.2) |
| 5 | 80 (0.7) | 73 (0.6) | 153 (0.7) |
| PCR, median (IQR) | 17.0 (9.2–40.5) | 14.0 (8.15–30.0) | 14.0 (8.5–31.5) |
| ACR, median (IQR) | 1.9 (0.7–5.5) | 2.1 (1.0–5.7) | 2.0 (0.8–5.6) |
| Comorbidities | |||
| Hypertension, % | 9723 (83.1) | 9654 (82.9) | 19,377 (83.0) |
| Systolic BP, mm Hg, mean (SD) | 134.9 (16.7) | 133.6 (16.0) | |
| Diastolic BP, mm Hg, mean (SD) | 74.9 (10.2) | 74.7 (9.9) | |
| Diabetes mellitus, n (%) | 2284 (19.5) | 1936 (16.6) | 4220 (18.1) |
| Cardiovascular disease, n (%) | 4786 (40.9) | 4750 (40.8) | 9536 (40.8) |
IQR, interquartile range; BMI, body mass index; 95% CI, 95% confidence interval; MDRD, Modification of Diet in Renal Disease; PCR, urine protein:creatinine ratio; ACR, urine albumin:creatinine ratio.
The PSP-CKD study employed 2.7 whole time–equivalent nurse practitioners to deliver the intervention, and practices were allocated equally between them pro rata. By the end of the study, the 23 practices in the intervention group had received a total of 984 visits (range, 23–62) from a CKD nurse. The mean number of visits per practice per year was 12.2, and the mean duration of each visit was 180.6 min (SD=77.7). Control group practices received 184 visits in total, but these were of short duration and restricted to data extraction activities only.
Most patients had CKD stage 3a, and the proportions with CKD stages 3–5 were similar in both control and intervention groups. The mean average eGFR per practice at baseline in the control group was 53.6 (95% confidence interval [95% CI], 52.7 to 54.5) ml/min per 1.73 m2, and in the intervention group, it was 54.6 (95% CI, 53.6 to 55.5) ml/min per 1.73 m2. After 42 months of follow-up, mean average eGFR change in the intervention group was −2.29 (95% CI, −3.89 to −0.69) ml/min per 1.73 m2, and in the control group, it was −2.00 (95% CI, −3.95 to 0.92) ml/min per 1.73 m2. There was no statistically significant difference in eGFR between the control and treatment groups at the end of the follow-up period or any other point in the study.
Secondary analyses of the primary outcome sequentially at 6-month intervals up to 42 months (Table 2) also revealed no significant differences between groups. Similarly, subgroup analyses showed no difference in the rate of decline of eGFR in patients with CKD stage 3a (control, −3.79; 95% CI, −6.31 to −1.28 versus intervention, −1.47; 95% CI, −3.07 to 0.12 ml/min per 1.73 m2) or 3b (control, 0.76; 95% CI, −1.18 to −2.69 versus intervention, −1.18; 95% CI, −2.46 to 0.09 ml/min per 1.73 m2) or all patients with CKD with a negative eGFR gradient at baseline (control, −2.28; 95% CI, −5.10 to 0.54 versus intervention, −2.09; 95% CI, −3.22 to −0.95 ml/min per 1.73 m2).
Table 2.
Secondary analysis of primary outcome over the course of the study
| Time, mo | Control | Intervention | P Value |
|---|---|---|---|
| 6 | −0.12 (−0.54 to 0.29) | −0.18 (−0.25 to 0.45) | 0.82 |
| 12 | 0.19 (−0.37 to 0.75) | 0.10 (−0.19 to 0.59) | 0.79 |
| 18 | −0.55 (−1.10 to −0.01) | −1.02 (−1.38 to −0.66) | 0.16 |
| 24 | −2.00 (−4.04 to 0.05) | −2.29 (−3.89 to −0.69) | 0.59 |
| 30 | −1.17 (−1.87 to −0.47) | −1.43 (−1.43 to −0.87) | 0.56 |
| 36 | −1.43 (−2.07 to −0.80) | −1.47 (−1.98 to −0.96) | 0.93 |
| 42 | −2.00 (−3.95 to 0.92) | −2.29 (−3.89 to −0.69) | 0.82 |
Values represent mean change in eGFR in milliliters per minute per 1.73 m2 (95% confidence interval), and a negative value represents a fall in eGFR.
Mean BP in both groups was similar at baseline (Table 1). After 36 months, the mean average BP in control practices was 128.7 (125.9–130.2)/73.7 (73.0–74.5), and in intervention practices, it was 130.2 (128.1–132.4)/73.8 (73.1–74.4; P=0.37). The number of patients achieving applicable BP targets was suboptimal in both groups at baseline. However, during the study, significantly more patients reached BP targets in the intervention group than in the control group at most time points (Table 3). As the study progressed, the proportion of patients reaching BP targets increased steadily in both groups, but it increased more quickly in the intervention group, where significantly more patients achieved BP targets at most time points.
Table 3.
Number of patients with CKD achieving BP control to target
| Time Point | Control | Intervention | Incidence Rate Ratio (95% CI) | |||
|---|---|---|---|---|---|---|
| Unadjusted | P Value | Adjusted | P Value | |||
| Baseline | 6834/11,551 (59.2) | 7137/11,473 (62.2) | ||||
| 6 mo | 6801/11,428 (59.5) | 7006/11,259 (60.7) | 1.05 (1.01 to 1.08) | <0.01 | 1.04 (1.01 to 1.08) | 0.02 |
| 12 mo | 6856/11,174 (61.4) | 7364/11,373 (64.7) | 1.06 (1.02 to 1.09) | 0.001 | 1.05 (1.02 to 1.09) | <0.01 |
| 18 mo | 6777/10,979 (61.7) | 7341/11,250 (65.2) | 1.06 (1.02 to 1.09) | 0.001 | 1.05 (1.02 to 1.09) | <0.01 |
| 24 mo | 6921/10,778 (64.2) | 7275/11,023 (66.0) | 1.03 (0.99 to 1.06) | 0.10 | 1.02 (0.99 to 1.06) | 0.18 |
| 30 mo | 6586/10,488 (62.8) | 7128/10,735 (66.4) | 1.06 (1.02 to 1.09) | 0.001 | 1.05 (1.02 to 1.09) | 0.003 |
| 36 mo | 6711/10,169 (66.0) | 7076/10,392 (68.1) | 1.03 (1.00 to 1.07) | 0.07 | 1.03 (1.00 to 1.07) | 0.06 |
| 42 mo | 7460/11,129 (67.0) | 7295/10,689 (68.2) | 1.02 (0.99 to 1.05) | 0.28 | 1.02 (0.99 to 1.06) | 0.22 |
Values represent numbers meeting target per total patients with CKD (percentage of total). BP values were imputed using the last observation carried forward whenever no measurements were taken within a time period. If a patient died or moved to a different trial practice with the opposing treatment allocation, they were excluded from the analysis thereafter. 95% CI, 95% confidence interval.
There were no significant differences in mean proteinuria values between groups measured as either ACR or PCR at baseline or after 42 months of the intervention. Similarly, in subgroup analyses of patients with CKD stages 3a and 3b, there were no differences in proteinuria values between groups before or after the intervention. Nonetheless, clinical coding for proteinuria increased significantly more in the intervention group throughout the study compared with the control group, and this difference was evident in both patients with CKD stage 3a and patients with CKD 3b (Table 4).
Table 4.
Number of patients with CKD and a proteinuria code in practice electronic records at sequential time points in both control and intervention groups
| Time Point | Control | Intervention | P Value |
|---|---|---|---|
| Baseline | 627/11,689 (5.4) | 982/11,546 (8.5) | |
| 6 mo | 1774/11,516 (15.4) | 1888/11,289 (16.7) | 0.01 |
| 12 mo | 1781/11,269 (15.8) | 2924/11,414 (16.8) | 0.06 |
| 18 mo | 1819/11,084 (16.4) | 1979/11,284 (17.5) | 0.04 |
| 24 mo | 1736/10,884 (16.0) | 1909/11,070 (17.2) | 0.02 |
| 30 mo | 1733/10,565 (16.4) | 1936/10,786 (17.9) | <0.01 |
| 36 mo | 1724/10,251 (16.8) | 1931/10,460 (18.5) | <0.01 |
| 42 mo | 1947/11,228 (17.3) | 2110/10,755 (19.6) | <0.001 |
Values represent n (%).
At the commencement of the study, many patients did not have a diagnostic code for CKD in their electronic practice record. There was a significantly greater increase in correctly coded patients with CKD in the intervention practices compared with the control practices such that, at 36 months in control practices, 32.5% of patients remained uncoded, whereas in intervention practices, 23.1% remained uncoded (Table 5).
Table 5.
Coding of patients with CKD in control and intervention practices
| Time Point | Control, n (%) | Intervention, n (%) | OR Adjusted for Baseline (95% CI) | P Value | ||
|---|---|---|---|---|---|---|
| Baseline | 11,706 | (100.0) | 11,651 | (100.0) | ||
| Correct | 6785 | (58.0) | 7268 | (62.4) | — | — |
| Not coded | 4313 | (36.8) | 3719 | (31.9) | — | — |
| 6 mo | 12,160 | (100.0) | 12,215 | (100.0) | ||
| Correct | 7069 | (58.1) | 7778 | (63.7) | 1.94 (1.09 to 3.46) | 0.03 |
| Not coded | 4461 | (36.7) | 3742 | (30.6) | 0.60 (0.39 to 0.91) | 0.02 |
| 12 mo | 12,568 | (100.0) | 13,312 | (100.0) | ||
| Correct | 7317 | (58.2) | 8595 | (64.6) | 1.94 (1.17 to 3.22) | 0.01 |
| Not coded | 4590 | (36.5) | 3963 | (29.8) | 0.55 (0.36 to 0.85) | <0.01 |
| 18 mo | 12,985 | (100.0) | 13,494 | (100.0) | ||
| Correct | 7808 | (60.1) | 8818 | (65.3) | 1.55 (0.94 to 2.58) | 0.09 |
| Not coded | 4486 | (34.5) | 3890 | (28.8) | 0.63 (0.39 to 1.01) | 0.05 |
| 24 mo | 13,398 | (100.0) | 13,855 | (100.0) | ||
| Correct | 8067 | (60.2) | 9239 | (66.7) | 1.80 (1.07 to 3.01) | 0.03 |
| Not coded | 4605 | (34.4) | 3798 | (27.4) | 0.54 (0.33 to 0.88) | 0.01 |
| 30 mo | 13,413 | (100.0) | 14,116 | (100.0) | ||
| Correct | 8070 | (60.2) | 9661 | (68.4) | 1.93 (1.19 to 3.14) | <0.01 |
| Not coded | 4584 | (34.2) | 3621 | (25.7) | 0.51 (0.31 to 0.81) | <0.01 |
| 36 mo | 13,760 | (100.0) | 14,179 | (100.0) | ||
| Correct | 8475 | (61.6) | 10,046 | (70.9) | 1.88 (1.22 to 2.90) | 0.004 |
| Not coded | 4473 | (32.5) | 3275 | (23.1) | 0.51 (0.33 to 0.79) | 0.002 |
| 42 mo | 6997 | (100.0) | 11,686 | (100.0) | ||
| Correct | 4398 | (62.9) | 8238 | (70.5) | 1.84 (1.01 to 3.36) | 0.05 |
| Not coded | 2143 | (30.6) | 2698 | (23.1) | 0.56 (0.28 to 1.10) | 0.09 |
Not coded patients met the definition of CKD on the basis of eGFR criteria but were not coded as such in clinical information technology systems. Correctly coded patients with CKD had been identified on the basis of eGFR criteria and had been correctly coded as such. OR, odds ratio; 95% CI, 95% confidence interval; —, not calculated.
A full-time CKD specialist nurse practitioner for ten practices would cost £56,442 per year, the cost of a primary care practice nurse including on costs and overheads but excluding qualifications and capital building costs.19 We estimate that one nurse could “reach” 576 patients with CKD across these ten practices, calculated using data from East Leicestershire and Rutland (average list size of 6000, of whom 0.96% have coded or uncoded CKD and uncontrolled BP). The annual cost per patient “reached” is, therefore, approximately £100, equivalent to around four general practitioner appointments.19 The cost per practice visit would be approximately £460 on the basis of the mean of 12.2 visits per practice per year.
Discussion
The PSP-CKD study is the largest study to evaluate the outcomes of nurse practitioner–led targeted CKD management in primary care. The results indicate that the intervention did not protect renal function but that it did have a significant positive and potentially cost-effective benefit on several important process measures and quality indicators of CKD care.
Several studies of multidisciplinary approaches to CKD care have been reported, mostly of small size and/or in secondary care. The MASTERPLAN study examined a secondary care nurse–led intervention in more advanced CKD where patients had a mean eGFR of 35 ml/min per 1.73 m2, and it found no effect on CV events but a 20% reduction in renal end points, although only after nearly 6 years of follow-up.20,21 The QICKD study found that audit-based education in primary care resulted in a modest reduction of BP in patients with CKD.22 Overall, however, there are few studies of CKD management interventions in primary care, and taken together, the reported efficacy of such interventions is minimal.23,24 In this context, the PSP-CKD study introduces important new findings. Although not affecting the progression of renal impairment in these patients with generally early CKD, significant positive effects of the PSP-CKD study intervention were noted on several process measures affecting key quality indicators in CKD care.
The PSP-CKD study is the largest ever intervention study in primary care CKD, and it has some significant strengths. In the United Kingdom, individual general practices are independent contractors, each with bespoke operations and IT systems. It was, therefore, essential that the PSP-CKD study could interrogate all IT systems to identify and capture data for all patients with CKD and that the study intervention was cost neutral to practices and structured to ensure acceptable integration into the operations of multiple distinct clinical environments. Hence, IMPAKT was successfully configured to be compatible with any IT system, and although study nurses were given facilitated access to practices, individual practices were free to choose to access an intensity of intervention most appropriate for their individual circumstances. Because this study design encouraged >50% of all practices in Northamptonshire to enroll in the PSP-CKD study and because intervention practices received around 3000 hours of the nurse intervention, this chosen approach in primary care was clearly broadly acceptable. In this regard, the senior-led initial engagement program was crucially important to gain confidence of general practitioners and their management teams and establish the clinical and academic credibility of the PSP-CKD study. Deployment of IMPAKT on practice IT systems served as a focal point for the intervention; this and the highly relevant audit data presented to practice staff by IMPAKT served as excellent tools to maintain practice staff engagement during the study.
Nonetheless, some practices declined to participate in the PSP-CKD study. Some research-naïve practices had never hosted research studies previously and were reluctant to participate. Other stated reasons given included overwhelming clinical pressures, key staff shortages, concomitant construction works in the practice, and doubts about the importance of CKD as a clinical priority.
The study also has some weaknesses. When the PSP-CKD study was designed, there were limited data describing the trajectory of eGFR change in unreferred primary care cohorts of patients with CKD, but some studies reported a rate of decline between 2.5 and 4.0 ml/min per 1.73 m2 per annum.25,26 Although O’Hare et al.27 found that eGFR declined at >3 ml/min per 1.73 m2 per annum in around 28% of approximately 175,000 patients with CKD stage 3 and around 35% of approximately 16,000 patients with CKD stage 4, elderly patients tended to lose renal function more slowly than younger patients. Supported by IMPAKT, the PSP-CKD study now provides robust evidence of the rate of change in eGFR in patients with CKD in the primary care setting. The study population in the PSP-CKD study was predominantly elderly, and in the PSP-CKD study, the rate of decline of eGFR was considerably lower than previously reported at <1 ml/min per 1.73 m2 per annum. Some practices also failed to complete 3.5 years of follow-up. Thus, the study was probably underpowered to detect its specified primary outcome. However, it seems unlikely that the primary outcome target difference in eGFR would have been met even if more clusters had been recruited into a much larger study. Indeed, even in subgroups at higher risk of renal function loss, no signal of an intervention effect was observed. Nonetheless, in retrospect, it should be acknowledged that seeking to identify a change in CKD progression in a population with only very slowly progressive disease was overly optimistic, and a negative result is unsurprising in this context. However, the absence of an effect of the intervention on eGFR progression in primary care remains an important negative observation.
The pragmatic design of the intervention meant that the intensity of application across clusters was not uniform. A more mandated, rigid application of the intervention may have led to a greater effect on measured outcomes but at the expense of reduced cluster recruitment and generalizability of findings for future rollout.
The PSP-CKD study results emphasize that CKD progression in primary care populations is generally slow; that, in many general practices, patients with CKD are not well identified; and that there is considerable potential for strengthening of processes of CKD care. Although targeted interventions to protect renal function may benefit some, for most patients with CKD, targeting CV risk is likely to be more important. The PSP-CKD study intervention significantly improved several CKD process measures, and these rather than outcome measures may be the most appropriate instrument for assessing quality in health care.28 One rationale for quality standards in CKD care is to mitigate the elevated risk of CV morbidity and mortality. Accordingly, NICE quality standards specify that adults with CKD should have regular eGFR and urine protein testing and that they should have their BP maintained within a recommended range.29 Elevated BP is a modifiable risk factor for both CV disease and CKD,30 and in CKD, these risks are exacerbated by proteinuria.31 Proteinuria is also a significant modifiable risk factor associated with more severe and progressive CKD.32 In the PSP-CKD study, a beneficial effect of the nurse intervention was observed on process measures relating to identifying and registering patients with CKD, assessing and coding proteinuria, and treating BP to target. Consequently, the intervention should have a beneficial effect on quality of CKD care, thus reducing CV morbidity and mortality with potential accompanying health economic benefits. The cost-effectiveness of proteinuria screening remains somewhat controversial. However, recent data using a Markov model to simulate the natural course of proteinuria-based disease progression to dialysis and occurrence of CV events yielded a favorable cost-effectiveness analysis.33 Most commentators endorse screening of proteinuria in high-risk populations, such as those with eGFR<60 ml/min.34
The cost of the CKD specialist nurse was low per patient “reached” such that small changes in outcomes could still be worthwhile, which has been shown for policy and service interventions in general.35 To provide a basic “headroom” analysis,36 we used the weighted mean cost of treating stroke and angina from the NHS Reference Costs,37 weighted by activity numbers, of £2067 to estimate that the intervention would be cost saving if it could prevent 28 such CV events: 1 per 21 patients reached. Weighted average utility weights for a CV event in the last 12 months of 0.6238 and CKD stages 3/4 of 0.7339 suggest a QALY saving of 0.11 per CV event averted. At a value of £20,000 per QALY, the number of CV events needed to be prevented for the intervention to be cost effective reduces to 13 or 1 per 43 patients reached. These estimates assume that the intervention is otherwise cost neutral when additional general practitioner visits may be required, although savings could be made elsewhere: for example, in reduced hospital appointment costs.
Guideline-based CKD management led by nurses in general practice is recommended to cost effectively deliver increased quality of care; a specific focus should be on improving processes that reduce CV risk. The interventions needed to achieve this rely on adequate data, but they are not complex and could be delivered in a care bundle alongside that for other long-term conditions.
These results provide the first robust evidence that a nurse-led, guideline-based intervention in primary care produces significant improvements in CKD care quality, and there could be accompanying health economic benefits.
Disclosures
Dr. Brown reports grants from National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care–East Midlands/West Midlands during the conduct of the study. Dr. Rogers reports nonfinancial support from National Health Service Northamptonshire outside the submitted work. Dr. Brunskill reports grants from the United Kingdom National Institutes of Health Research during the conduct of the study. All of the remaining authors have nothing to disclose.
Supplementary Material
Acknowledgments
The Primary-Secondary Care Partnership to Prevent Adverse Outcomes in Chronic Kidney Disease study was funded initially by the Collaboration for Leadership in Applied Health Research and Care (CLAHRC) Leicestershire Northamptonshire and Rutland and then, by CLAHRC–East Midlands. Dr. Brown is supported by CLAHRC–West Midlands. The study was subject to external peer review by the National Institute for Health Research CLAHRC Scientific Committee before funding approval.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
Supplemental Material
This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2018101042/-/DCSupplemental.
Supplemental Material. PSP-CKD study protocol with sample size amendment.
References
- 1.Stevens PE, O’Donoghue DJ, de Lusignan S, Van Vlymen J, Klebe B, Middleton R, et al.: Chronic kidney disease management in the United Kingdom: NEOERICA project results. Kidney Int 72: 92–99, 2007 [DOI] [PubMed] [Google Scholar]
- 2.Major R, Shepherd D, Warwick G, Brunskill N: Prescription rates of cardiovascular medications in a large UK primary care chronic kidney disease cohort. Nephron 133: 15–22, 2016 [DOI] [PubMed] [Google Scholar]
- 3.Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de Jong PE, et al.: Chronic Kidney Disease Prognosis Consortium : Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: A collaborative meta-analysis. Lancet 375: 2073–2081, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fox CS, Matsushita K, Woodward M, Bilo HJ, Chalmers J, Heerspink HJ, et al.: Chronic Kidney Disease Prognosis Consortium : Associations of kidney disease measures with mortality and end-stage renal disease in individuals with and without diabetes: A meta-analysis. Lancet 380: 1662–1673, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Stevens LA, Coresh J, Greene T, Levey AS: Assessing kidney function--measured and estimated glomerular filtration rate. N Engl J Med 354: 2473–2483, 2006 [DOI] [PubMed] [Google Scholar]
- 6.National Institute for Health and Care Excellence: Chronic Kidney Disease in Adults: Assessment and Management, 2014. Available at: https://www.nice.org.uk/guidance/cg182. Accessed April 26, 2018 [PubMed]
- 7.Levin A, Stevens PE, Bilous RW, Coresh J, De Francisco ALM, De Jong PE, et al. : Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl 3: 1–50, 2013 [Google Scholar]
- 8.Department of Health: National Service Framework for Renal Services - Part Two: Chronic Kidney Disease, Acute Renal Failure and End of Life Care, 2005. Available at: http://webarchive.nationalarchives.gov.uk/20130124065433/http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/@en/documents/digitalasset/dh_4102680.pdf. Accessed April 26, 2018
- 9.NHS Employers: Quality and Outcomes Framework Guidance for GMS Contract 2008/09. Delivering Investment in General Practice, 2008. Available at: http://www.nhsemployers.org/-/media/Employers/Documents/Primary-care-contracts/QOF/2008-09/Quality-and-Outcomes-Framework-guidance-for-GMS-contract-200809—Delivering-investments-in-general-practice.pdf. Accessed April 26, 2018
- 10.Crinson I, Gallagher H, Thomas N, de Lusignan S: How ready is general practice to improve quality in chronic kidney disease? A diagnostic analysis. Br J Gen Pract 60: 403–409, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Moynihan R, Glassock R, Doust J: Chronic kidney disease controversy: How expanding definitions are unnecessarily labelling many people as diseased. BMJ 347: f4298, 2013 [DOI] [PubMed] [Google Scholar]
- 12.Black C, Sharma P, Scotland G, McCullough K, McGurn D, Robertson L, et al.: Early referral strategies for management of people with markers of renal disease: A systematic review of the evidence of clinical effectiveness, cost-effectiveness and economic analysis. Health Technol Assess 14: 1–184, 2010 [DOI] [PubMed] [Google Scholar]
- 13.Hall YN, Himmelfarb J: The CKD classification system in the precision medicine era. Clin J Am Soc Nephrol 12: 346–348, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Allen AS, Forman JP, Orav EJ, Bates DW, Denker BM, Sequist TD: Primary care management of chronic kidney disease. J Gen Intern Med 26: 386–392, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Inrig JK, Califf RM, Tasneem A, Vegunta RK, Molina C, Stanifer JW, et al.: The landscape of clinical trials in nephrology: A systematic review of Clinicaltrials.gov. Am J Kidney Dis 63: 771–780, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ruggenenti P, Perna A, Loriga G, Ganeva M, Ene-Iordache B, Turturro M, et al.: REIN-2 Study Group : Blood-pressure control for renoprotection in patients with non-diabetic chronic renal disease (REIN-2): Multicentre, randomised controlled trial. Lancet 365: 939–946, 2005 [DOI] [PubMed] [Google Scholar]
- 17.Strippoli GF, Bonifati C, Craig M, Navaneethan SD, Craig JC: Angiotensin converting enzyme inhibitors and angiotensin II receptor antagonists for preventing the progression of diabetic kidney disease. Cochrane Database Syst Rev 4: CD006257, 2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Campbell MK, Fayers PM, Grimshaw JM: Determinants of the intracluster correlation coefficient in cluster randomized trials: The case of implementation research. Clin Trials 2: 99–107, 2005 [DOI] [PubMed] [Google Scholar]
- 19.Curtis L, Burns A: Unit Costs of Health and Social Care 2016, Personal Social Services Research Unit, Canterbury, United Kingdom, University of Kent, 2016 [Google Scholar]
- 20.van Zuilen AD, Bots ML, Dulger A, van der Tweel I, van Buren M, Ten Dam MA, et al.: Multifactorial intervention with nurse practitioners does not change cardiovascular outcomes in patients with chronic kidney disease. Kidney Int 82: 710–717, 2012 [DOI] [PubMed] [Google Scholar]
- 21.Peeters MJ, van Zuilen AD, van den Brand JA, Bots ML, van Buren M, Ten Dam MA, et al.: Nurse practitioner care improves renal outcome in patients with CKD. J Am Soc Nephrol 25: 390–398, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lusignan S, Gallagher H, Jones S, Chan T, van Vlymen J, Tahir A, et al.: Audit-based education lowers systolic blood pressure in chronic kidney disease: The Quality Improvement in CKD (QICKD) trial results. Kidney Int 84: 609–620, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Galbraith L, Jacobs C, Hemmelgarn BR, Donald M, Manns BJ, Jun M: Chronic disease management interventions for people with chronic kidney disease in primary care: A systematic review and meta-analysis. Nephrol Dial Transplant 33: 112–121, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bello AK, Qarni B, Samimi A, Okel J, Chatterley T, Okpechi IG, et al.: Effectiveness of multifaceted care approach on adverse clinical outcomes in nondiabetic CKD: A systematic review and meta-analysis. Kidney Int Rep 2: 617–625, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Richards N, Harris K, Whitfield M, O’Donoghue D, Lewis R, Mansell M, et al.: Primary care-based disease management of chronic kidney disease (CKD), based on estimated glomerular filtration rate (eGFR) reporting, improves patient outcomes. Nephrol Dial Transplant 23: 549–555, 2008 [DOI] [PubMed] [Google Scholar]
- 26.Bayliss EA, Bhardwaja B, Ross C, Beck A, Lanese DM: Multidisciplinary team care may slow the rate of decline in renal function. Clin J Am Soc Nephrol 6: 704–710, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.O’Hare AM, Choi AI, Bertenthal D, Bacchetti P, Garg AX, Kaufman JS, et al.: Age affects outcomes in chronic kidney disease. J Am Soc Nephrol 18: 2758–2765, 2007 [DOI] [PubMed] [Google Scholar]
- 28.Lilford RJ, Brown CA, Nicholl J: Use of process measures to monitor the quality of clinical practice. BMJ 335: 648–650, 2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.National Institute for Health and Care Excellence: Chronic Kidney Disease in Adults. NICE Quality Standard 5, 2011. Available at: https://www.nice.org.uk/guidance/qs5/resources/chronic-kidney-disease-in-adults-pdf-58297746373. Accessed April 26, 2018
- 30.Hsu CY, McCulloch CE, Darbinian J, Go AS, Iribarren C: Elevated blood pressure and risk of end-stage renal disease in subjects without baseline kidney disease. Arch Intern Med 165: 923–928, 2005 [DOI] [PubMed] [Google Scholar]
- 31.Peterson JC, Adler S, Burkart JM, Greene T, Hebert LA, Hunsicker LG, et al.: Blood pressure control, proteinuria, and the progression of renal disease. The modification of diet in renal disease study. Ann Intern Med 123: 754–762, 1995 [DOI] [PubMed] [Google Scholar]
- 32.Gansevoort RT, Correa-Rotter R, Hemmelgarn BR, Jafar TH, Heerspink HJ, Mann JF, et al.: Chronic kidney disease and cardiovascular risk: Epidemiology, mechanisms, and prevention. Lancet 382: 339–352, 2013 [DOI] [PubMed] [Google Scholar]
- 33.Boersma C, Gansevoort RT, Pechlivanoglou P, Visser ST, van Toly FF, de Jong-van den Berg LT, et al.: Prevention of Renal and Vascular End Stage Disease Study Group : Screen-and-treat strategies for albuminuria to prevent cardiovascular and renal disease: Cost-effectiveness of nationwide and targeted interventions based on analysis of cohort data from The Netherlands. Clin Ther 32: 1103–1121, 2010 [DOI] [PubMed] [Google Scholar]
- 34.Berns JS: Routine screening for CKD should be done in asymptomatic adults... selectively. Clin J Am Soc Nephrol 9: 1988–1992, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lilford RJ, Chilton PJ, Hemming K, Girling AJ, Taylor CA, Barach P: Evaluating policy and service interventions: Framework to guide selection and interpretation of study end points. BMJ 341: c4413, 2010 [DOI] [PubMed] [Google Scholar]
- 36.Cosh E, Girling A, Lilford R, McAteer H, Young T: Investing in new medical technologies: A decision framework. J Commer Biotechnol 13: 263–271, 2007 [Google Scholar]
- 37.Department of Health : National Schedule of Reference Costs 2014-15, London, 2015. Available at: https://www.gov.uk/government/publications/nhs-reference-costs-2014-to-2015. Accessed
- 38.Ara R, Brazier JE: Populating an economic model with health state utility values: Moving toward better practice. Value Health 13: 509–518, 2010 [DOI] [PubMed] [Google Scholar]
- 39.Wyld M, Morton RL, Hayen A, Howard K, Webster AC: A systematic review and meta-analysis of utility-based quality of life in chronic kidney disease treatments. PLoS Med 9: e1001307, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.