INTRODUCTION

Chronic kidney disease (CKD), defined as a decrease in renal function (glomerular filtration rate [GFR] or creatinine clearance rate <60ml/min/173m2) or as persistent renal damage lasting at least three months,1,2 constitutes an important health problem in industrialised nations. CKD is associated with an aging population, arterial hypertension (AHT), and diabetes mellitus (DM).3 The prevalence of this disease increases with age. According to the results of the EPIRCE study, the prevalence of CKD in Spain is 9.16%, a number that increases to 23.7% in individuals older than 64 years of age.4 CKD also makes individuals 3-10 times more at risk for cardiovascular events as compared to reference populations,5,6 and is associated with increased mortality and hospitalisation rates.7,8 Early stages of CKD tend to be asymptomatic and easily detectable in the primary care (PC) setting during routine check-ups in older patients, those with hypertension, and those with diabetes, by determining plasma creatinine levels, GFR as estimated using standard formulas,2,3 and by calculating urine albumin/creatinine ratios in a simple urine sample. However, these tests are not always applied, such that a large portion of patients have hidden CKD; that is to say, reduced GFR but with plasma creatinine values that fall within the normal range in laboratory tests.3 The importance of early detection of CKD is based on the fact that early actions to combat factors that favour the progression of kidney disease (such as proteinuria, AHT, and DM), cessation of potentially nephrotoxic medications, and optimising drug doses for GFR may slow the progression of renal failure and its associated complications.2

Given the high prevalence of CKD, its economic repercussions (in Spain, the cost of treating this disease in its advanced stages is estimated at 800 million Euros/year),2 the associated risk for cardiovascular disease, and the availability of medications that are effective at impeding or slowing its progression,9 clinical guidelines have been developed for the diagnosis and proper follow-up of cases of CKD directed towards both PC doctors and nephrologists (K/DOQI1, KDIGO,10 S.E.N.,11 CMJA,12 NICE,13 SIGN,14 Renal Association,15 among others). Experts and health authorities such as the World Health Organisation (WHO) have alerted the medical community as to the need for effectively disseminating the information in these documents in order to facilitate appropriate medical practices. However, the level of awareness and implementation of these evidence-based clinical recommendations is insufficient.

In 2007, the Spanish Society of Family and Community Medicine (semFYC) and the Spanish Society of Nephrology (S.E.N.) created a consensus document with the goal of decreasing the variability in clinical practices for the detection, treatment, and referral of patients with CKD, and to promote health care collaboration and coordination between PC and nephrology.2 This S.E.N.-semFYC consensus document on chronic kidney disease included 13 recommendations based on scientific evidence covering the following topics: 1) optimisation of treatment; 2) urinary pathology; 3) patient screening; 4) evaluation of renal function; 5) estimating GFR using equations; 6) estimating GFR using other methods; 7) evaluating urinary excretion of proteins; 8) vascular risk (detection and management) 9) controlling vascular risk factors; 10) prevention of iatrogenic events; 11) referral to nephrology; 12) referral of diabetic patients to nephrology; and 13) PC-nephrology collaboration.

The objective of this study was to evaluate the level of acceptance, awareness, agreement, and implementation of these guidelines among Spanish nephrologists and PC physicians two years after publication.

MATERIAL AND METHOD

Ours was an observational, descriptive, and cross-sectional study carried out through the application of a self-administered, voluntary, written survey between May and November 2009. The questionnaire was composed of 24 items regarding the recommendations made in the S.E.N.-semFYC consensus document, grouped into four sections: the concept of CKD, detection and diagnosis of CKD, therapeutic approach, and referral to nephrology. Each of the items comprised two multiple choice questions with answers referring to the appropriateness and use of the S.E.N.-semFYC recommendations evaluated in each item (Appendix I).

The questionnaires were distributed along with a letter of invitation to both medical societies (nephrology and primary care) by mail, with a second dispatch if necessary, to 550 family doctors and 200 nephrologists, of which 326 PC physicians and 150 specialists responded. The total number of doctors participating in the survey was then stratified into sub-samples corresponding to the proportion of the population treated in each autonomous community. The sample of randomly selected participants was large and geographically diverse, yielding a high estimated precision for the responses to the survey items, with a maximum error of ±5% for PC and ±8% for nephrology, with a 95% confidence level.

The eligibility criteria for participation in the study were: practicing nephrologists and PC physicians of any sex or age, working at public and private Spanish health care institutions anywhere within the country, and who frequently treat patients with AHT, diabetes, and/or CKD.

Variables and measurements

The initial block of questions in the survey referred to the profile of each participant (demographic data, professional experience, field of medicine, type of patients treated, workload, etc.). The second block of questions was a questionnaire designed to evaluate the participant’s level of awareness, agreement, and implementation in daily practice of the clinical recommendations regarding CKD contained within the S.E.N.-semFYC consensus document on chronic kidney disease (Table 1). The questionnaire ended with questions concerning the opinion of each participant regarding the usefulness and need for this document.

Members of each professional society also responded to a group of questions considered by the scientific committee of the study to be of particular interest for each speciality: PC (items 1-11 and 13-18) and nephrology (items 4-5, 8-12, and 15-24). Some of these items were designed exclusively for only one speciality (PC: 1-3, 6-7, 13, and 14; nephrology: 12 and 19-24), whereas others were responded to by members of both specialities.

For each clinical recommendation evaluated, different aspects were assessed based on the speciality of the survey participant. PC physicians were asked for their personal evaluation regarding prior awareness of the recommendation and the level of implementation in their daily practice. Nephrologists were asked to give a professional evaluation of the recommendation and to state their level of implementation of each recommendation based on personal criteria.

Once completed, the paper questionnaires were scanned by trained personnel using mrScan® software, an application that facilitates automatic preparation of printed questionnaires for scanning and digitalising the results. This obviated the need for double manual entry of information.

Preparation, analysis, and interpretation of the results

The survey results were tabulated in terms of relative frequency distributions (%) for each response option for each item and for each medical speciality (PC and nephrology).

We analysed the significance of differences between PC physicians and nephrologists in self-reported awareness regarding each recommendation using chi-square tests.

In order to identify possible differences in the global level of agreement expressed by survey participants on an ordinal scale, we obtained mean and median values for the score reported for each item (categorising these responses on an ordinal scale of 1 to 5), and constructed a 95% confidence interval around the mean. We analysed the significance of differences in the criteria/opinion expressed by members of each speciality using Mann-Whitney non-parametric U-tests.

We also used standard statistical tests (chi-square, Student’s t, analysis of variance, or their non-parametric equivalents) to analyse possible associations between personal and professional characteristics of the survey participants (medical speciality, age, field of medicine, professional experience, etc.) and certain characteristics of the test.

Ethical/legal aspects

Our study was a survey of professional attitudes and opinions that did not affect routine clinical practice or produce any type of intervention on patients. The study also did not compile any information regarding patient identities or their pathologies. The data compiled regarding professional attitudes and opinions were processed anonymously, and results were not linked in any way to the identity of study participants. Data regarding researchers were treated based on organic law 15/1999, from 13 December, for the protection of personal information.

RESULTS

Professional environment and participant profiles

We surveyed a total of 476 doctors from all Spanish autonomous communities. The profiles of the study participants are summarised in Table 2. Compared to nephrologists, PC physicians had a longer history of medical practice and attended to more patients/week with cardiovascular risk, defined as patients >60 years of age, with AHT, or with DM. Nephrologists, on the other hand, tended to treat patients in urban areas and treated more patients with CKD than PC physicians.

Nephrologists had a greater level of access to GFR estimates using the MDRD formula (Modification of Diet in Renal Disease, MDRD-GFR) in their normal laboratory analyses (86.8% vs 66.6% in PC physicians; P<.001). There were no differences in terms of creatinine clearance estimates using the Cockcroft-Gault formula (72.1% in PC physicians vs 64.6% in nephrologists) or albumin/creatinine ratios (Alb/Cr) in urine samples (94.4% in nephrologists vs 94.1% in PC physicians).

None of the following descriptive variables of the professional profiles of the study participants (age, sex, public or private practice, professional experience, geographical area, number of patients with CKD or risk factors treated, or access to diagnostic tests of renal function in laboratory analyses) were associated with the level of awareness or perceived need/usefulness of the consensus document.

Level of awareness of the S.E.N.-semFYC consensus document on chronic kidney disease in primary care

Only 51.1% of PC physicians reported familiarity with the contents of the S.E.N.-semFYC consensus document before participation in the survey.

The level of awareness among PC physicians of each of the 24 items in the questionnaire regarding the consensus document (Table 1) that were selected specifically for PC physicians (1-11, 13-18) varied widely: between 0.9% and 50% of PC physicians surveyed reported that they had no previous familiarity with these characteristics. In 5 of the 24 items (5, 9, 10, 15, and 16a), the percentage of prior unfamiliarity surpassed 20% of survey participants (Table 3), this constituting poor awareness of the recommendation. A level of unfamiliarity of 10%-20% was considered to be normal, and 0%-10% was considered to be indicative of good awareness. These five items of “poor awareness” among PC physicians are particularly relevant, since these include the criteria for the appropriateness of urinary ultrasound analyses in males older than 60 years of age with CKD, and the criteria for referring patients to nephrology.

Level of agreement with the S.E.N.-semFYC consensus document on chronic kidney disease among nephrologists

The level of professional agreement (consideration of the document as appropriate) with each of the recommendations made as manifested by the nephrologists surveyed reached a high level in all cases (total or partial agreement) (Figure 1).

The mean percentage of nephrologists in complete agreement with the specific recommendations for their speciality (“total agreement”) was 64.1% (range: 34.7%-83.1%), and the level of partial agreement (“tends to agree”) was 29.7% (range: 16.9%-52%).

The recommendation that yielded the greatest level of agreement was item 20: “In addition to the aforementioned criteria, diabetic patients should be referred to nephrology when also suffering from AHT refractory to treatment (three different drugs at maximum doses and still no control)”, with an 83.1% rate of total agreement among nephrologists surveyed.

We also observed a strong correlation between the level of professional agreement reported for each recommendation and the level of implementation of said recommendation in routine clinical practice (Spearman’s coefficient >0.4).

Level of implementation of the S.E.N.-semFYC consensus document on chronic kidney disease in nephrologists and primary care physicians

The level of implementation in routine clinical practice reported for the recommendations in the consensus document that were specific for nephrologists was very high; in all cases (Figure 2), the mean score was situated between “always” (51.2%) and “almost always” (38%). Only 0.7% and 8.1% of nephrologists answered “never” or “rarely”, respectively. The recommendation corresponding to item 12: “Albumin/creatinine ratio provides a good estimate of proteinuria and obviates the need for taking 24-hour urine samples” was the least closely followed recommendation of all, with 8.1% of survey participants reporting “never” or “rarely”.

Upon comparing the level of implementation of recommendations between PC physicians and nephrologists, only 2 items (11 and 15, regarding the evaluation of urinary protein excretion and referral to nephrology, respectively) were equally followed by nephrologists and PC physicians. For all other items, nephrologists reported a greater level of implementation than PC physicians; “always”: nephrologists: 51.2% (range: 27.4%-74.8%) vs. PC physicians: 35.8% (range: 22.7%-72.3%); “almost always”: nephrologists: 38% (range: 22.4%-49.3%) vs PC physicians: 42.6% (range: 23.9%-54.4%). This difference was especially notable for three items (Table 4): item 5, regarding obstructive pathologies in male patients >60 years of age (93.6% vs 60.4%); item 9, regarding creatinine clearance measurements in 24-hour urine samples for estimating GFR (80.4% vs 64.5%), and item 10, regarding the numerical results of GFR (85% vs 44.1%).

Perception of the need and usefulness of the S.E.N.-semFYC consensus document on chronic kidney disease

As a group, nephrologists manifested a greater perceived need for the S.E.N.-semFYC consensus document on chronic kidney disease in order to improve the quality of clinical practice. However, the usefulness of the document for clinical practice was given greater value among PC physicians than among nephrologists (P<.001) (Figure 3).

DISCUSSION

The results from our study provide useful information regarding the profiles of doctors who treat CKD patients, their level of familiarity with this pathology, and their opinions regarding the recommendations and level of application of the S.E.N.-semFYC consensus document on chronic kidney disease.

CKD has a high prevalence in the general population, which places the PC physician at an ideal position for identifying the at-risk population and providing early detection and treatment of this disease.16 Referral of patients to nephrology should be limited to select cases, primarily those with the greatest risk for progression or already in advanced stages of CKD. In fact, some programmes for shared treatment of CKD between PC and nephrology reveal that a substantial number of renal patients do not require referral.17

One of the most notable results from our survey is that, despite a major effort for dissemination throughout all implicated scientific societies, only half of the PC physicians that responded to the questionnaire were familiar with the S.E.N.-semFYC consensus document on chronic kidney disease. These results are similar to those published regarding awareness of CKD clinical practice guidelines from other countries.18-20

We also observed that PC physicians are not very familiar with the recommendations regarding patient referral to nephrology, and these are the items that are least applied in clinical practice. Two of every three PC physicians surveyed reported that they always or almost always apply recommendation 15 (no need for referring patients >70 years of age with stable CKD in stage 1-3 and albuminuria <500mg/g). Taking into account the large number of CKD patients that fall under these criteria (in the EPIRCE study,4 52% of all patients with CKD were older than 65 years of age with stage 1-3 CKD), lack of awareness and implementation of this recommendation among PC physicians is an important point for improvement.

The results of our survey reflect a very positive opinion among nephrologists regarding these guidelines, with 70.8% of those surveyed reporting that they considered the guidelines to be useful and at least moderately influential in daily clinical practice (52.1%). These results are similar to those from surveys carried out in other countries,21,22 and reflect the potential capacity of evidence-based documents for modifying the clinical decision making process.22

The survey demonstrated that 30% of PC physicians and 20% of nephrologists do not normally use the CKD stage classification system provided in the consensus document (the same as proposed by the KDIGO and National Kidney Foundation guidelines). This may be due to the fact that insufficient time has passed since the publication of the consensus document and guidelines for CKD in order to completely modify the clinical practice of the health care professionals surveyed. This is compounded by the variability in the definitions used for CKD in original publications from the medical literature. In a recent review, the internationally accepted definition of CKD was specifically mentioned and clearly adopted in only 20% of all articles on CKD published in 2009.23

As regards the laboratory analyses used to evaluate renal function, the Spanish  Society of Clinical Biochemistry and Molecular Pathology, together with the S.E.N., established recommendations in 200724 specifying that glomerular filtration rates “must be estimated using equations based on serum creatinine levels, thus obviating the need for measuring creatinine clearance in 24-hour urine samples” in clinical practice. However, our survey showed that almost 30% of PC physicians did not recognise that “creatinine clearance as measured in 24-hour urine samples does not generally improve the accuracy of GFR values as obtained using equations”, which highlights the need for implementing strategies for disseminating these recommendations among Spanish doctors.

At the time of the survey, the normally used equations for estimating glomerular filtration rate (Cockcroft-Gault and MDRD) were well-known and routinely used by PC physicians (practically 85% were familiar with the equations and more than 75% were in complete agreement with their use). The most recent equation, the CKD-EPI, was not well established at the time of the survey, and so it was not included in the questionnaire.

We observed a high level of professional agreement among nephrologists with the recommendations outlined in the consensus document, as was the case for the level of implementation of these recommendations, except for the item regarding urinary albumin excretion: “albumin/creatinine ratios provide a good estimate of proteinuria and obviate the need for taking 24-hour urine samples”, in which the level of implementation was lower. Three different tests are commonly used in the medical literature for measuring proteinuria: dipstick, protein/creatinine ratio in urine samples, and 24-hour urine samples,25 although current guidelines preferably recommend the use of the albumin/creatinine ratio.26 The reference intervals for Alb/Cr do not take into account differences in creatinine excretion between different populations of patients (related to differences in sex, age, and ethnicity) or the increase in risk related to albumin excretion,27 which could explain why this is the least commonly applied of all recommendations from the S.E.N.-semFYC consensus document on chronic kidney disease by nephrologists (8.1% reported that they “never” or “almost never” followed this recommendation).

The level of implementation appears to be correlated with prior familiarity as reported by the survey participants. Although PC physicians declared a substantial level of implementation of the consensus document recommendations, this was significantly lower than the high level of implementation reported by nephrologists. This could be due to the fact that, compared with nephrologists, PC physicians use much more heterogeneous criteria for evaluating patients with CKD.28

This is the first cross-sectional study to evaluate the opinions of both PC physicians and nephrologists regarding guidelines and recommendations that are widely distributed among practitioners of both specialities. Our results show that the implementation of clinical practice guidelines, in this case regarding CKD, is not a simple process due to the presence of barriers for development and application. We have shown that the S.E.N.-semFYC consensus document would have the potential to modify the clinical practice of health care professionals if the dissemination of these recommendations among PC physicians were greater. The information provided by our study could provide a strong baseline upon which to develop future training measures for the health care professionals involved in the global process of treating patients with CKD (both nursing staff and medical residents), primarily in the form of prevention. Our study has also pointed out the need for improving the awareness and use of this document among PC physicians, with special emphasis on the criteria for referring patients to nephrology.

Acknowledgements

We would like to thank all family doctors and nephrologists who participated in this study by completing an anonymous and impartial survey, which produced the information used to elaborate this manuscript.

Conflicts of interest

The authors declare the following potential conflicts of interest: this study was sponsored and supported by the S.E.N. and the semFYC. We also received logistical support from Luzan Editores (Research Division). The project was funded by Chiesi laboratories.

11367_19157_39288_en_w477712109113ref.1136717858_11367_16025_26453_es_11367_tabla1.doc

Table 1. Content of the context recommendations evaluated in our study

11367_19157_39289_en_w477712109114ref.1136717858_11367_19115_37398_es_11367_tabla2.doc

Table 2. Profiles of the doctors surveyed

11367_19157_39290_en_w477712109115ref.1136717858_11367_19115_37400_es_11367_tabla3_copy11.doc

Table 3. Level of awareness of the consensus recommendations as reported by primary care physicians surveyed

11367_19157_39291_en_w477712109116ref.1136717858_11367_19115_37411_es_11367_tabla4_copy2.doc

Table 4. Difference in the level of implementation of consensus recommendations reported by physicians of both specialties

11367_19157_39292_en_w477712109117ref.1136717858_11367_19115_26454_es_11367_figura1.doc

Figure 1. Level of agreement with the S.E.N-semFYC guidelines, nephrologists

11367_19157_39293_en_w477712109118ref.1136717858_11367_19115_37409_es_11367_figura2.doc

Figure 2. Level of implementation of the S.E.N.-semFYC consensus guidelines among nephrologists and primary care physicians

11367_19157_39294_en_w477712109119ref.1136717858_11367_19115_37414_es_11367_figura3.doc

Figure 3. Need for the consensus document and usefulness for improving clinical practices for treating chronic kidney disease

11367_19157_39295_en_11367_anexo_1_en.doc

Survey-evaluation of the S.E.N.-semFYC document