J Nephrol DOI 10.1007/s40620-015-0175-3
ORIGINAL ARTICLE
Deficits in information transfer between hospital-based and primary-care physicians, the case of kidney disease: a cross-sectional study Be´ne´dicte Sautenet • Agne`s Caille • Bruno Giraudeau Julie Le´ger • Patrick Vourc’h • Matthias Buchler • Jean-Michel Halimi
•
Received: 31 October 2014 / Accepted: 22 January 2015 Italian Society of Nephrology 2015
Abstract Background Late recognition plays an important role in prognosis associated with kidney disease; thus, information transfer at hospital discharge regarding kidney disease is crucial. Whether it is notified in patients’ hospital discharge summary (HDS) is presently largely unknown. Study design Cross-sectional. Setting and participants The prevalence of kidney dysfunction [estimated glomerular filtration rate (eGFR) \60 ml/min/1.73 m2] and its reporting to primary-care physicians from 26 units [11 surgery, 11 medical, 4 intensive care units (ICUs)] of a university hospital were analyzed in 14,000 hospitalizations. Predictor eGFR. Outcome Notification of kidney dysfunction in HDS. Measurements GFR was estimated from serum creatinine using the Modification of Diet in Renal Disease formula.
B. Sautenet (&) M. Buchler J.-M. Halimi Service de Ne´phrologie-Immunologie Clinique, CHU de Tours, Hoˆpital Bretonneau, 37044 Tours, France e-mail: [email protected] B. Sautenet A. Caille B. Giraudeau J. Le´ger Centre d’Investigation Clinique INSERM, CIC 1415, Tours, France B. Sautenet A. Caille B. Giraudeau M. Buchler J.-M. Halimi Universite´ Franc¸ois-Rabelais, Tours, France P. Vourc’h CHU de Tours, Laboratoire de Biochimie et Biologie Mole´culaire, Tours, France M. Buchler J.-M. Halimi EA 4245, Universite´ Franc¸ois-Rabelais, Tours, France
Results Kidney dysfunction was frequent (27.2 %) but infrequently notified in the main-body of the HDS (overall 25.3 %, medical 25 %, surgical 16.3 %, ICU 48.4 %) even when severe (eGFR 15–29.9 ml/min/1.73 m2 was notified in 68.8, 38.5, and 79.8 % of HDSs in medical, surgical and ICUs, respectively). Notification in the HDS conclusion was rare (overall 11.4 %, medical 9.8 %, surgical 8.4 %, ICU 27.5 %). Reporting remained low when eGFR remained abnormal at discharge (medical 35.8 %, surgical 22.5 %, ICU 62.2 %) but was worse for acute kidney injury (16.0, 17.1, and 37.7 %, respectively). The optimal eGFR cut-off for reporting was 39 ml/min/1.73 m2. Longer durations of hospitalization, greater numbers of creatinine measurements and of abnormal eGFR were associated with notification, regardless of the type of unit. Limitations Lack of data to define acute or chronic kidney injury with precision. Conclusions Kidney dysfunction is frequent in hospitalized patients but is usually not notified, even when severe or still present at discharge, suggesting that it is not considered important to disclose to primary-care physicians. This lack of information may decrease physicians’ awareness, and may affect continuity of care in patients with kidney dysfunction. Keywords Information Primary-care physician Hospital discharge summary Kidney disease
Introduction Continuity of patient care after hospital discharge requires communication and information to primary care physicians. Incomplete documentation at hospital discharge has substantial implications for continuity of care, patient
123
J Nephrol
safety, patient and clinician satisfaction and resource use [1]. In a recent review of 73 studies, the availability of a discharge summary at the first post-discharge visit was low (12–34 %), and affected the quality of care in approximately 25 % of follow-up visits [1]. Moreover, discharge summaries often lacked important information such as diagnostic test results (missing in 33–63 % of cases), treatment or hospital course (7–22 %), discharge medications (2–40 %), test results pending at discharge (65 %), patient or family counseling (90–92 %), and follow-up plans (2–43 %) [1]. Serum creatinine is routinely measured to estimate glomerular filtration rate (eGFR), which allows easy detection of acute kidney injury (AKI) and chronic kidney disease (CKD). In a recent study, AKI incidence was 22.7 % among 19,249 hospitalizations, and the mortality rate was 10.8 % for patients with AKI versus 1.5 % for those without [2]. In another study, AKI was associated with the development of CKD, end-stage renal disease (ESRD) and death [3]. Moreover, CKD is associated with an increased risk of death, cardiovascular events, and hospitalization [4–6]. Therefore, both AKI and CKD are associated with morbidity and mortality [2–4]. For these reasons, primary-care physicians should be informed of the occurrence of AKI or CKD in their patients. This information is critical for continuity of primary care, including eventual referral to a nephrologist. Unfortunately, referral to a nephrologist is often late, probably because physicians lack this information [7]. In 2007, late referral was associated with an increased risk of death among 12,749 patients [8]. In contrast, early referral determined a significantly shorter initial hospitalization and reduced 5-year mortality in a recent review of 27 cohort studies [9]. Finally, early recognition of kidney disease is essential to improve prognosis. This early recognition and management of AKI or CKD occurring in the hospital requires timely communication and information to primary-care physicians [10, 11]. The communication and information regarding kidney dysfunction, how this information is transmitted to primarycare physicians and how the presence of AKI or CKD is taken into account in the hospital discharge summary (HDS) is largely unknown. In the present study, we analyzed the prevalence of kidney dysfunction (AKI or CKD) and its notification to primary-care physicians in a university hospital.
Patients and methods Selection of HDS and assessment of kidney dysfunction All serum creatinine measurements made in hospitalized adult patients at the Tours University Hospital between
123
November 1, 2008 and April 30, 2009 were selected to perform this cross-sectional study. All serum creatinine values were recorded for each hospitalization. Glomerular filtration rate was estimated with the simplified Modification of Diet in Renal Disease (MDRD) formula, and we recorded whether eGFR was \60 ml/min/1.73 m2 [10, 12]. For each hospitalization, we analyzed all eGFR values. We defined whether kidney dysfunction was persistent (i.e. last eGFR value at discharge \60 ml/min/1.73 m2 when serum creatinine was measured several times during the same hospitalization) or whether it was transient (i.e. AKI). Twenty-six units were included: • •
•
4 intensive care units (ICUs) (medical ICU, surgical ICU, neurology ICU and burns unit); 11 surgical units (cardiovascular, orthopedic, urology, abdominal, neurology, obstetric-gynecology, ear–nose– throat, chest, ophthalmology, plastic and maxillofacial surgery); 11 medical units (internal medicine, cardiology, rheumatology, dermatology, neurology, pulmonary medicine, gastroenterology, oncology, hematology, geriatric medicine, and infectious diseases).
The HDS was excluded from the analysis if the patient: (1) died during hospitalization; (2) was admitted to a nephrology unit; (3) was admitted to a psychiatry unit (because serum creatinine is not routinely measured, and patients hospitalized in psychiatry units in Tours University Hospital do not receive an HDS); or (4) was admitted to a pediatric unit and under 18 years of age (because the MDRD formula does not accurately estimate GFR in children). Analysis of the main body and conclusion of HDS regarding kidney dysfunction during hospitalization To identify the HDSs in which kidney dysfunction was reported, a key-words list was established: ‘‘GFR’’, ‘‘MDRD’’, ‘‘Cockcroft’’, ‘‘creatinine clearance’’, ‘‘nephrologist’’, ‘‘nephrology’’, ‘‘nephrologic’’, ‘‘nephropathy’’, ‘‘kidney dysfunction’’, ‘‘kidney ultrasound’’, ‘‘dialysis’’, ‘‘hemodialysis’’, ‘‘chronic kidney disease’’, ‘‘acute kidney injury’’ and ‘‘anuric’’. All HDSs with at least one of these keywords were read and assessed by one nephrologist (BS). To assess the negative predictive value of this extraction technique, 200 random HDS without these key-words were analyzed: no case of kidney dysfunction was reported, which validated our selection process. When a patient was hospitalized several times during the study period, all HDSs were analyzed. When a patient was admitted to several units, the HDS of each unit was analyzed. Referral to a nephrologist during the hospitalization and planned referral after discharge were noted.
J Nephrol Fig. 1 Flow chart showing selection process of hospital discharge c summaries
We evaluated whether the information regarding kidney dysfunction appeared in the main body or in the conclusion of the HDS. When kidney dysfunction was indicated in the conclusion of the HDS, we interpreted this as a clear intention to communicate this information to the primary-care physician. Statistical analysis Quantitative data are presented as mean and standard deviation (SD) for normally distributed variables and median, first and third quartiles for variables with skewed distributions. Qualitative data are described with percentages. Comparisons were made using the Chi square test for qualitative data and Wilcoxon test for quantitative data. To further clarify the association of the staging of kidney dysfunction and its reporting in the HDS, we used receiver operating characteristic (ROC) curve analysis and calculated an area under the curve (AUC): the minimal eGFR threshold for maximal sensitivity and specificity was assessed by the Youden index. A p value \0.05 was considered as significant.
Results Prevalence of kidney dysfunction during hospitalization Between November 1, 2008 and April 30, 2009, there were 19,798 hospitalizations in the Tours University Hospital; serum creatinine was measured in 14,000/19,798 (70.7 %) hospitalizations, corresponding to 98.1 % (999/1,018), 90.6 % (8,578/9,468), and 47.5 % (4,423/9,312) of the hospitalizations in ICU, medical, and surgical units respectively. Overall, kidney dysfunction (defined as eGFR \60 ml/ min/1.73 m2 on one or more occasions) was observed in 27.2 % (3,801/14,000) (Fig. 1). These 3,801 hospitalizations corresponded to 3,342 patients (mean age 74.4 ± 13.4 years; men 50.8 %). The prevalence of kidney dysfunction was 27.0 % in medical units, 24.9 % in surgical units and 38.2 % in ICUs (Table 1); the prevalence of moderate kidney dysfunction (i.e. eGFR between 45 and 59 ml/min/1.73 m2) was 15.6 % in medical units, 14.7 % in surgical units and 15.4 % in ICUs (Fig. 2). Reporting of kidney dysfunction in HDS Influence of type of unit Hospital discharge summary was present for 3,421/3,801 hospitalizations (medical units 93.6 %, surgical units
123
J Nephrol Table 1 Prevalence of kidney dysfunction in 14,000 hospitalizations Medical units (n = 8,578)
Surgical units (n = 4,423)
ICUs (n = 999)
Kidney dysfunction, n (%)
Yes, 2,318 (27.0)
No, 6,260 (73.0)
Yes, 1,101 (24.9)
No, 3,322 (75.1)
Yes, 382 (38.2)
Age, median (Q1–Q3) (years)
78 (68–85)
61 (48–74)
76 (66–83)
59 (44–73)
70 (59–79)
No, 617 (61.8) 52 (36–68)
Sex, n (%) (male)
1,131 (48.8)
3,585 (57.3)
600 (54.5)
1,756 (52.9)
70 (54.7)
205 (62.5)
Length of hospitalization, median (Q1–Q3) (days)
4 (2–10)
3 (2–7)
7 (3–12)
6 (3–10)
16 (8–28)
8 (3–18)
Kidney dysfunction was defined as an eGFR value \60 ml/min/1.73 m2 eGFR estimated glomerular filtration rate (ml/min/1.73 m2), ICU intensive care unit, Q1–Q3 interquartile range
Fig. 2 Prevalence of kidney dysfunction in medical units, surgical units and intensive care units
79.7 %, ICUs 97.9 %). Overall, kidney dysfunction was reported in 25.3 % of HDSs. Kidney dysfunction was reported more frequently in ICUs (48.4 %) than in medical (25.0 %) or surgical (16.3 %) units (Table 2). Reports of kidney dysfunction were associated with longer median duration of stay [medical units 7 (4–14) vs. 4 (2–8) days; surgery units: 9 (7–16) vs. 7 (4–12) days; ICUs: 19 (11–30) vs. 14 (7–25) days, p \ 0.001], a greater number of serum creatinine measurements [medical units: 3 (1–5) vs. 2 (1–3); surgery units: 7 (3–11) vs. 2 (1–5) days; ICUs: 11 (7–21) vs. 9 (5–14), p \ 0.001] and a greater number of eGFR values \60 ml/min/1.73 m2 [medical units: 2 (1–4) vs. 1 (1, 2); surgery units: 4 (2–9) vs. 1 (1–3); ICU: 8 (3–14) vs. 2 (1–5), p \ 0.001]. Results did not differ among the different units (Table 2). Reporting of kidney dysfunction: main body vs. conclusion of HDS
combination were noted in 2.5 % of HDSs). Referral to a nephrologist (during hospitalization or planned referral after discharge) was indicated in 1.7 % of HDSs (Table 3). Influence of eGFR levels Among medical units, kidney dysfunction was reported in 10.3, 31.3, 68.8 and 89.5 % of HDSs for eGFR values 45–59, 30–44, 15–29 and \15 ml/min/1.73 m2, respectively. Similar results were observed for all units (Table 2). Using ROC curves, we found that the optimal cut-off for reporting kidney dysfunction in HDSs was 39 ml/min/ 1.73 m2 (AUC 0.81, sensitivity 0.69, specificity 0.80) in medical units; 40 ml/min/1.73 m2 (0.79, 0.71, and 0.76, respectively) in surgical units, and 36 ml/min/1.73 m2 (0.86, 0.81, and 0.83, respectively) in ICUs. Influence of duration of kidney dysfunction
In medical units, kidney dysfunction was reported in the conclusion of the HDS in only 11.4 % hospitalizations (medical units 9.8 %; surgical units 8.4 %; ICUs 27.5 %) (Table 2). AKI was noted in 7.0 % of HDSs, whereas CKD was noted in 13.1 % (both AKI and CKD in
123
When eGFR remained \60 ml/min/1.73 m2 at discharge, kidney dysfunction was reported in 35.8 % of HDSs in medical units, in 22.5 % in surgical units and in 62.2 % in ICUs (Table 4). The proportions were much lower when
49 (41–55) 49 (41–55) 32 (18–43) 33 (23–44) 50 (41–55)
5 (2–9)
3 (2–7)
31 (20–43)
Number of creatinine measurements, median (Q1–Q3)
Number eGFR values \60 ml/min/1.73 m2, median (Q1–Q3)
Lowest eGFR value, median (Q1–Q3)
2
10 (5–20) Duration of hospitalization, median (Q1–Q3) (days)
84.7 eGFR 0–14 ml/min/1.73 m2
Parameters associated with reporting of kidney dysfunction
65.3 eGFR 15–29 ml/min/1.73 m2
389 (11.4)
30.2 eGFR 30–44 ml/min/1.73 m2
Notification in conclusion of HDS, n (%)
9.7 eGFR 45–59 ml/min/1.73 m2
Reporting by eGFR value:
Yes, 866 (25.3)
HDS hospital discharge summary, eGFR estimated glomerular filtration rate (ml/min/1.73 m2), ICU intensive care unit, Q1–Q3 interquartile range
1 (1–3) 4 (2–9) 1 (1, 2) 2 (1–4)
Influence of age class
Kidney dysfunction was defined as a eGFR value \60 ml/min/1.73 m . All comparisons by units were significant (p \ 0.001)
2 (1–5) 7 (3–11)
1 (1–3)
kidney dysfunction was due to AKI (16.0, 17.1, and 37.7 % in medical, surgical units and ICUs, respectively).
23 (13–32)
19 (11–30) 7 (4–12) 9 (7–16) 4 (2–8)
2 (1–3) 3 (1–5)
7 (4–14) 6 (2–12)
2 (1–5)
212 (9.8)
89.5
50 (42–55)
103 (27.5) 74 (8.4)
9 (5–14)
91.4 61.7
2 (1–5)
79.8 38.5 68.8
8 (3–14)
43.9 22.3 31.3
10.3
Yes, 542 (25.0) No, 2,555 (74.7)
11 (7–20)
16.7 6.1
Yes, 181 (48.4) Yes, 143 (16.3) No, 1,627 (75.0)
No, 735 (83.7)
ICUs (n = 374)
Kidney dysfunction reported, n (%)
Table 2 Reporting of kidney dysfunction in HDSs
All units (n = 3,421)
Medical units (n = 2,169)
Surgical units (n = 878)
No, 193 (51.6)
14 (7–25)
J Nephrol
We divided patients into 3 age classes: 180 patients were under 50 years, 1,860 were 50–79 years, and 1,381 were over 80 years old. The reporting of kidney dysfunction according to age class and subdivided by level of eGFR and hospitalization unit is presented in Table 5. Low eGFR tended to be more frequently reported in patients aged \50 years. However, in patients with eGFR between 30 and 44 ml/min/1.73 m2, similar figures were observed regardless of age (Table 5). Overall, reporting was less than 30 % among patients aged \50 years (Table 5).
Discussion The results of our study indicate that kidney dysfunction frequently occurs during hospitalization, but is noted in the HDS in only one out of four patients. The rate of reporting of kidney dysfunction may vary by type of hospitalization unit (medical, surgical, or ICU), eGFR level, and duration of renal dysfunction (AKI vs. persisting renal dysfunction or CKD) but it remained low. The intention to convey information is reflected by the content of the HDS conclusion: the reporting of kidney dysfunction was present in only 11.4 % of HDS conclusions. Our study is one of the few assessing how kidney dysfunction is reported to primary-care physicians via the HDS. In 2007, a review of communication with the primarycare physicians at hospital discharge found that results of diagnostic tests were not mentioned in 33–63 % of HDSs, but the specific diagnostic tests missing were not detailed [1]. Our study instead gives complete information on the transmission of kidney dysfunction information at patient discharge via the HDS. The lack of knowledge of CKD has been investigated in different situations. In 2008, Rothberg et al. [13] showed that 46 % of CKD cases in an elderly population were recognized by physicians. In 2009, Gentile et al. [14] found that diagnostic codes were correctly used for only 19 % of hospitalized patients with CKD in an Italian hospital. The same year, Guessous et al. [15] estimated in a North American health system study that only 14.4 % (1,478/ 10,266) of patients with CKD were diagnosed. The lack of knowledge of CKD implies a delay in management of patients with increased hospitalizations and mortality [10]. Our study confirms the lack of knowledge regarding kidney dysfunction, and provides an analysis by type of hospitalization unit (medical, surgical and ICU). To improve the quality of care of patients with CKD, several institutions
123
J Nephrol Table 3 Kidney dysfunction reporting: specific information included in the HDS Units
All units (3,421)
Medical (2,169)
Surgical (878)
Intensive care (374)
AKI noted
239 (7.0)
145 (6.7)
87 (9.9)
137 (36.6)
CKD noted
449 (13.1)
336 (15.5)
44 (5.0)
69 (18.4)
Superimposed AKI on CKD
84 (2.5)
39 (1.8)
14 (1.6)
31 (8.3)
Planned referral to a nephrologist
59 (1.7)
45 (2.1)
3 (0.3)
11 (2.9)
Data are number (%) HDS hospital discharge summary, AKI acute kidney injury, CKD chronic kidney disease
Table 4 Reporting of kidney dysfunction in the main body and conclusion of the HDS according to the type of kidney dysfunction (persistent kidney dysfunction vs. acute kidney injury) Units
All units (n = 2,328)
Medical (n = 1,314)
Surgical (n = 651)
Intensive care (n = 363)
Kidney dysfunction reported
Yes
p value
Yes
p value
Yes
p value
Yes
p value
514 (35.6)
\0.001
334 (35.8)
\0.001
78 (22.5)
0.087
102 (62.2)
\0.001
In the main body of HDS, n (%) Persistent kidney dysfunction at discharge Acute kidney injury In the conclusion of HDS, n (%)
188 (21.2)
Persistent kidney dysfunction at discharge
236 (16.4)
Acute kidney injury
61 (16.0) 0.854
130 (13.9)
88 (9.9)
52 (17.1) 0.673
22 (5.8)
36 (10.4) 36 (11.8)
75 (37.7) 0.051
70 (42.7)
\0.001
30 (15.1)
Of note, hospitalizations were selected when serum creatinine was measured more than once HDS hospital discharge summary
Table 5 Reporting of kidney dysfunction in the HDS according to age class
Kidney dysfunction reported, n (%) eGFR (ml/min/1.73 m2)
30–44 (n = 935)
15–29 (n = 406)
28/99 (28.3)
15/41 (36.6)
16/21 (76.2)
Medical units
16/46 (34.8)
10/14 (71.4)
4/7 (57.1)
3/3 (100.0)
Surgical units
3/27 (11.1)
2/4 (50.0)
2/4 (50.0)
5/7 (71.4)
Intensive care units
9/26 (34.6)
3/3 (100.0)
10/10 (100.0)
7/9 (77.8)
\50 years (n = 180) All units
45–59 (n = 1,918)
0–14 (n = 162)
15/19 (78.9)
50–79 years (n = 1,860) All units
177/1,153 (15.4)
164/444 (36.9)
122/171 (71.3)
78/92 (84.8)
Medical units
127/724 (17.5)
106/275 (38.5)
72/96 (75.0)
21/24 (87.5)
Surgical units
24/333 (7.2)
28/110 (25.5)
12/31 (38.7)
18/29 (62.1)
30/59 (50.8)
38/44 (86.4)
39/39 (100.0)
Intensive care units
26/96 (27.1)
[80 years (n = 1,381)
HDS hospital discharge summary, eGFR estimated glomerular filtration rate
All units
147/666 (22.1)
172/450 (38.2)
142/214 (66.4)
46/51 (90.2)
Medical units
132/478 (27.6)
140/320 (43.8)
102/141 (72.3)
28/30 (93.3)
Surgical units
11/35 (31.4)
24/114 (21.1)
18/43 (41.9)
8/11 (72.7)
4/28 (14.3)
8/16 (50.0)
22/30 (73.3)
10/10 (100.0)
Intensive care units
such as the US National Kidney Foundation recommended the systematic estimation of GFR in medical laboratories for each creatinine serum measurement [16]. In 2010, Hemmelgarn et al. [17] estimating the impact of systematic estimation of GFR showed an increase in nephrology consultations; however, no meaningful consequences were noted for patients. In 2012, Akbari et al. [18] showed an
123
increase in nephrology consultations after the systematic estimation of GFR but also an increase in inappropriate consultations. Patients with kidney dysfunction during hospitalization were seldom diagnosed, with disparity among units. Awareness of CKD has improved with the systematic estimation of GFR by laboratories. However, specific training
J Nephrol
of medical professionals to increase the rate of detection and adequate management of kidney dysfunction is also important. A study performed in 2004 by Akbari et al. [19] analyzed an educational program of primary-care providers and showed an improvement in detection of CKD. Overall, the lack of knowledge of CKD involves a delay in management of patients with increased hospitalizations and mortality, and missed opportunities to initiate appropriate treatment strategies. The training of hospital physicians to detect kidney disease and convey this diagnostic information to primary-care physicians seems an important step. Another step is the adequate training of primary-care physicians to manage patients with CKD within the framework of an educational public health program. We assessed eGFR during hospitalization but we had no data to define AKI or CKD with precision in the present study. Moreover, the severity indicated by the level of eGFR differs according to the patient’s age and comorbidities. The aim of our study was to describe the reporting of kidney dysfunction, not to ensure that the diagnosis of AKI or CKD was accurate, nor to assess the severity of kidney dysfunction. To estimate renal function, we used the MDRD formula rather than the calculation of creatinine clearance by the Cockcroft–Gault formula because bodyweight data was missing in most of the hospitalizations. Of note, GFR estimated by the MDRD formula is better in subjects in stable conditions, and the estimate may be biased in acutely ill patients; however, there is presently no adequate GFR formula in this situation [20]. We analyzed only hospitalizations with creatinine serum measurements. Given that more that 90 % of patients had a serum creatinine measurement in medical units and ICUs, the rate of kidney dysfunction we observed can be considered as the true prevalence in these units. Our study is a monocentric study performed in a university hospital: is it unknown whether communication regarding kidney dysfunction may be different in non-university hospitals. We did not analyze management of kidney dysfunction during hospitalizations: although this aspect is important, it was beyond the scope of our study. Importantly, systematic estimation of eGFR was not performed in our hospital, and thus it was not possible to assess whether systematic eGFR reporting could modify these findings. Whether automatic reporting of eGFR would modify the HDS reporting is presently unknown but warrants investigation. Kidney dysfunction is frequent but its reporting in the HDS is often missing, suggesting that this information is usually not considered important to disclose to primarycare physicians. The absence of this information in the HDS conclusions should be analyzed in other settings. The consequences of our observations should be studied in terms of continuity of patient care after discharge and late referral to nephrologists.
Conflict of interest of interest. Ethical standard required.
The authors declare that they have no conflict
For this type of study formal consent is not
References 1. Kripalani SLF (2007) Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA 297(8):831–841 2. Wang HE, Muntner P, Chertow GM, Warnock DG (2012) Acute kidney injury and mortality in hospitalized patients. Am J Nephrol 35(4):349–355 3. Coca SG, Singanamala S, Parikh CR (2012) Chronic kidney disease after acute kidney injury: a systematic review and metaanalysis. Kidney Int 81(5):442–448 4. James MT, Hemmelgarn BR, Tonelli M (2010) Early recognition and prevention of chronic kidney disease. Lancet 375(9722):1296–1309 5. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu C (2004) Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 351(13):1296–1305 6. Stevens LA, Fares G, Fleming J, Martin D, Murthy K, Stark PC, Uhlig K, Van Lente F, Levey AS (2005) Low rates of testing and diagnostic codes usage in a commercial clinical laboratory: evidence for lack of physician awareness of chronic kidney disease. J Am Soc Nephrol 16(8):2439–2448 7. Jungers P (2002) Late referral: loss of chance for the patient, loss of money for society. Nephrol Dial Transplant 17(3):371–375 8. Chan MR, Dall AT, Fletcher KE, Lu N, Trivedi H (2007) Outcomes in patients with chronic kidney disease referred late to nephrologists: a meta-analysis. Am J Med 120(12):1063–1070 9. Smart NA, Titus TT (2011) Outcomes of early versus late nephrology referral in chronic kidney disease: a systematic review. Am J Med 124(11):1073–1080.e2 10. Levey AS, Coresh J (2012) Chronic kidney disease. Lancet 379(9811):165–180 11. Kim DH, Kim M, Kim H, Kim YL, Kang CW, Kim NH, Kim YS, Lee JP (2013) Early referral to a nephrologist improved patient survival: prospective cohort study for end-stage renal disease in Korea. PLoS ONE 8(1):e55323 12. Palevsky PM, Liu KD, Brophy PD, Chawla LS, Parikh CR, Thakar CV, Tolwani AJ, Waikar SS, Weisbord SD (2013) KDOQI US commentary on the 2012 KDIGO clinical practice guideline for acute kidney injury. Am J Kidney Dis 61(5):649–672 13. Rothberg MB, Kehoe ED, Courtemanche AL, Kenosi T, Pekow PS, Brennan MJ, Mulhern JG, Braden GL (2008) Recognition and management of chronic kidney disease in an elderly ambulatory population. J Gen Intern Med 23(8):1125–1130 14. Gentile G, Postorino M, Mooring RD, De Angelis L, Manfreda VM, Ruffini F, Pioppo M, Quintaliani G (2009) Estimated GFR reporting is not sufficient to allow detection of chronic kidney disease in an Italian regional hospital. BMC Nephrol 10:24 15. Guessous I, McClellan W, Vupputuri S, Wasse H (2009) Low documentation of chronic kidney disease among high-risk patients in a managed care population: a retrospective cohort study. BMC Nephrol 10:25 16. National Kidney Foundation (2002) K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 39(Suppl 1):S1–S266 17. Hemmelgarn BR, Zhang J (2010) Nephrology visits and health care resource use before and after reporting estimated glomerular filtration rate. JAMA 303(12):1151–1158
123
J Nephrol 18. Akbari A, Grimshaw J, Stacey D, Hogg W, Ramsay T, ChengFitzpatrick M, Magner P, Bell R, Karpinski J (2012) Change in appropriate referrals to nephrologists after the introduction of automatic reporting of the estimated glomerular filtration rate. CMAJ 184(5):E269–E276 19. Akbari ASP (2004) Detection of chronic kidney disease with laboratory reporting of estimated glomerular filtration rate and an educational program. Arch Intern Med 164(16):1788–1792
123
20. Poggio ED, Nef PC, Wang X, Greene T, Van Lente F, Dennis VW, Hall PM (2005) Performance of the Cockcroft–Gault and modification of diet in renal disease equations in estimating GFR in ill hospitalized patients. Am J Kidney Dis 46(2):242–252
ORIGINAL ARTICLE
Deficits in information transfer between hospital-based and primary-care physicians, the case of kidney disease: a cross-sectional study Be´ne´dicte Sautenet • Agne`s Caille • Bruno Giraudeau Julie Le´ger • Patrick Vourc’h • Matthias Buchler • Jean-Michel Halimi
•
Received: 31 October 2014 / Accepted: 22 January 2015 Italian Society of Nephrology 2015
Abstract Background Late recognition plays an important role in prognosis associated with kidney disease; thus, information transfer at hospital discharge regarding kidney disease is crucial. Whether it is notified in patients’ hospital discharge summary (HDS) is presently largely unknown. Study design Cross-sectional. Setting and participants The prevalence of kidney dysfunction [estimated glomerular filtration rate (eGFR) \60 ml/min/1.73 m2] and its reporting to primary-care physicians from 26 units [11 surgery, 11 medical, 4 intensive care units (ICUs)] of a university hospital were analyzed in 14,000 hospitalizations. Predictor eGFR. Outcome Notification of kidney dysfunction in HDS. Measurements GFR was estimated from serum creatinine using the Modification of Diet in Renal Disease formula.
B. Sautenet (&) M. Buchler J.-M. Halimi Service de Ne´phrologie-Immunologie Clinique, CHU de Tours, Hoˆpital Bretonneau, 37044 Tours, France e-mail: [email protected] B. Sautenet A. Caille B. Giraudeau J. Le´ger Centre d’Investigation Clinique INSERM, CIC 1415, Tours, France B. Sautenet A. Caille B. Giraudeau M. Buchler J.-M. Halimi Universite´ Franc¸ois-Rabelais, Tours, France P. Vourc’h CHU de Tours, Laboratoire de Biochimie et Biologie Mole´culaire, Tours, France M. Buchler J.-M. Halimi EA 4245, Universite´ Franc¸ois-Rabelais, Tours, France
Results Kidney dysfunction was frequent (27.2 %) but infrequently notified in the main-body of the HDS (overall 25.3 %, medical 25 %, surgical 16.3 %, ICU 48.4 %) even when severe (eGFR 15–29.9 ml/min/1.73 m2 was notified in 68.8, 38.5, and 79.8 % of HDSs in medical, surgical and ICUs, respectively). Notification in the HDS conclusion was rare (overall 11.4 %, medical 9.8 %, surgical 8.4 %, ICU 27.5 %). Reporting remained low when eGFR remained abnormal at discharge (medical 35.8 %, surgical 22.5 %, ICU 62.2 %) but was worse for acute kidney injury (16.0, 17.1, and 37.7 %, respectively). The optimal eGFR cut-off for reporting was 39 ml/min/1.73 m2. Longer durations of hospitalization, greater numbers of creatinine measurements and of abnormal eGFR were associated with notification, regardless of the type of unit. Limitations Lack of data to define acute or chronic kidney injury with precision. Conclusions Kidney dysfunction is frequent in hospitalized patients but is usually not notified, even when severe or still present at discharge, suggesting that it is not considered important to disclose to primary-care physicians. This lack of information may decrease physicians’ awareness, and may affect continuity of care in patients with kidney dysfunction. Keywords Information Primary-care physician Hospital discharge summary Kidney disease
Introduction Continuity of patient care after hospital discharge requires communication and information to primary care physicians. Incomplete documentation at hospital discharge has substantial implications for continuity of care, patient
123
J Nephrol
safety, patient and clinician satisfaction and resource use [1]. In a recent review of 73 studies, the availability of a discharge summary at the first post-discharge visit was low (12–34 %), and affected the quality of care in approximately 25 % of follow-up visits [1]. Moreover, discharge summaries often lacked important information such as diagnostic test results (missing in 33–63 % of cases), treatment or hospital course (7–22 %), discharge medications (2–40 %), test results pending at discharge (65 %), patient or family counseling (90–92 %), and follow-up plans (2–43 %) [1]. Serum creatinine is routinely measured to estimate glomerular filtration rate (eGFR), which allows easy detection of acute kidney injury (AKI) and chronic kidney disease (CKD). In a recent study, AKI incidence was 22.7 % among 19,249 hospitalizations, and the mortality rate was 10.8 % for patients with AKI versus 1.5 % for those without [2]. In another study, AKI was associated with the development of CKD, end-stage renal disease (ESRD) and death [3]. Moreover, CKD is associated with an increased risk of death, cardiovascular events, and hospitalization [4–6]. Therefore, both AKI and CKD are associated with morbidity and mortality [2–4]. For these reasons, primary-care physicians should be informed of the occurrence of AKI or CKD in their patients. This information is critical for continuity of primary care, including eventual referral to a nephrologist. Unfortunately, referral to a nephrologist is often late, probably because physicians lack this information [7]. In 2007, late referral was associated with an increased risk of death among 12,749 patients [8]. In contrast, early referral determined a significantly shorter initial hospitalization and reduced 5-year mortality in a recent review of 27 cohort studies [9]. Finally, early recognition of kidney disease is essential to improve prognosis. This early recognition and management of AKI or CKD occurring in the hospital requires timely communication and information to primary-care physicians [10, 11]. The communication and information regarding kidney dysfunction, how this information is transmitted to primarycare physicians and how the presence of AKI or CKD is taken into account in the hospital discharge summary (HDS) is largely unknown. In the present study, we analyzed the prevalence of kidney dysfunction (AKI or CKD) and its notification to primary-care physicians in a university hospital.
Patients and methods Selection of HDS and assessment of kidney dysfunction All serum creatinine measurements made in hospitalized adult patients at the Tours University Hospital between
123
November 1, 2008 and April 30, 2009 were selected to perform this cross-sectional study. All serum creatinine values were recorded for each hospitalization. Glomerular filtration rate was estimated with the simplified Modification of Diet in Renal Disease (MDRD) formula, and we recorded whether eGFR was \60 ml/min/1.73 m2 [10, 12]. For each hospitalization, we analyzed all eGFR values. We defined whether kidney dysfunction was persistent (i.e. last eGFR value at discharge \60 ml/min/1.73 m2 when serum creatinine was measured several times during the same hospitalization) or whether it was transient (i.e. AKI). Twenty-six units were included: • •
•
4 intensive care units (ICUs) (medical ICU, surgical ICU, neurology ICU and burns unit); 11 surgical units (cardiovascular, orthopedic, urology, abdominal, neurology, obstetric-gynecology, ear–nose– throat, chest, ophthalmology, plastic and maxillofacial surgery); 11 medical units (internal medicine, cardiology, rheumatology, dermatology, neurology, pulmonary medicine, gastroenterology, oncology, hematology, geriatric medicine, and infectious diseases).
The HDS was excluded from the analysis if the patient: (1) died during hospitalization; (2) was admitted to a nephrology unit; (3) was admitted to a psychiatry unit (because serum creatinine is not routinely measured, and patients hospitalized in psychiatry units in Tours University Hospital do not receive an HDS); or (4) was admitted to a pediatric unit and under 18 years of age (because the MDRD formula does not accurately estimate GFR in children). Analysis of the main body and conclusion of HDS regarding kidney dysfunction during hospitalization To identify the HDSs in which kidney dysfunction was reported, a key-words list was established: ‘‘GFR’’, ‘‘MDRD’’, ‘‘Cockcroft’’, ‘‘creatinine clearance’’, ‘‘nephrologist’’, ‘‘nephrology’’, ‘‘nephrologic’’, ‘‘nephropathy’’, ‘‘kidney dysfunction’’, ‘‘kidney ultrasound’’, ‘‘dialysis’’, ‘‘hemodialysis’’, ‘‘chronic kidney disease’’, ‘‘acute kidney injury’’ and ‘‘anuric’’. All HDSs with at least one of these keywords were read and assessed by one nephrologist (BS). To assess the negative predictive value of this extraction technique, 200 random HDS without these key-words were analyzed: no case of kidney dysfunction was reported, which validated our selection process. When a patient was hospitalized several times during the study period, all HDSs were analyzed. When a patient was admitted to several units, the HDS of each unit was analyzed. Referral to a nephrologist during the hospitalization and planned referral after discharge were noted.
J Nephrol Fig. 1 Flow chart showing selection process of hospital discharge c summaries
We evaluated whether the information regarding kidney dysfunction appeared in the main body or in the conclusion of the HDS. When kidney dysfunction was indicated in the conclusion of the HDS, we interpreted this as a clear intention to communicate this information to the primary-care physician. Statistical analysis Quantitative data are presented as mean and standard deviation (SD) for normally distributed variables and median, first and third quartiles for variables with skewed distributions. Qualitative data are described with percentages. Comparisons were made using the Chi square test for qualitative data and Wilcoxon test for quantitative data. To further clarify the association of the staging of kidney dysfunction and its reporting in the HDS, we used receiver operating characteristic (ROC) curve analysis and calculated an area under the curve (AUC): the minimal eGFR threshold for maximal sensitivity and specificity was assessed by the Youden index. A p value \0.05 was considered as significant.
Results Prevalence of kidney dysfunction during hospitalization Between November 1, 2008 and April 30, 2009, there were 19,798 hospitalizations in the Tours University Hospital; serum creatinine was measured in 14,000/19,798 (70.7 %) hospitalizations, corresponding to 98.1 % (999/1,018), 90.6 % (8,578/9,468), and 47.5 % (4,423/9,312) of the hospitalizations in ICU, medical, and surgical units respectively. Overall, kidney dysfunction (defined as eGFR \60 ml/ min/1.73 m2 on one or more occasions) was observed in 27.2 % (3,801/14,000) (Fig. 1). These 3,801 hospitalizations corresponded to 3,342 patients (mean age 74.4 ± 13.4 years; men 50.8 %). The prevalence of kidney dysfunction was 27.0 % in medical units, 24.9 % in surgical units and 38.2 % in ICUs (Table 1); the prevalence of moderate kidney dysfunction (i.e. eGFR between 45 and 59 ml/min/1.73 m2) was 15.6 % in medical units, 14.7 % in surgical units and 15.4 % in ICUs (Fig. 2). Reporting of kidney dysfunction in HDS Influence of type of unit Hospital discharge summary was present for 3,421/3,801 hospitalizations (medical units 93.6 %, surgical units
123
J Nephrol Table 1 Prevalence of kidney dysfunction in 14,000 hospitalizations Medical units (n = 8,578)
Surgical units (n = 4,423)
ICUs (n = 999)
Kidney dysfunction, n (%)
Yes, 2,318 (27.0)
No, 6,260 (73.0)
Yes, 1,101 (24.9)
No, 3,322 (75.1)
Yes, 382 (38.2)
Age, median (Q1–Q3) (years)
78 (68–85)
61 (48–74)
76 (66–83)
59 (44–73)
70 (59–79)
No, 617 (61.8) 52 (36–68)
Sex, n (%) (male)
1,131 (48.8)
3,585 (57.3)
600 (54.5)
1,756 (52.9)
70 (54.7)
205 (62.5)
Length of hospitalization, median (Q1–Q3) (days)
4 (2–10)
3 (2–7)
7 (3–12)
6 (3–10)
16 (8–28)
8 (3–18)
Kidney dysfunction was defined as an eGFR value \60 ml/min/1.73 m2 eGFR estimated glomerular filtration rate (ml/min/1.73 m2), ICU intensive care unit, Q1–Q3 interquartile range
Fig. 2 Prevalence of kidney dysfunction in medical units, surgical units and intensive care units
79.7 %, ICUs 97.9 %). Overall, kidney dysfunction was reported in 25.3 % of HDSs. Kidney dysfunction was reported more frequently in ICUs (48.4 %) than in medical (25.0 %) or surgical (16.3 %) units (Table 2). Reports of kidney dysfunction were associated with longer median duration of stay [medical units 7 (4–14) vs. 4 (2–8) days; surgery units: 9 (7–16) vs. 7 (4–12) days; ICUs: 19 (11–30) vs. 14 (7–25) days, p \ 0.001], a greater number of serum creatinine measurements [medical units: 3 (1–5) vs. 2 (1–3); surgery units: 7 (3–11) vs. 2 (1–5) days; ICUs: 11 (7–21) vs. 9 (5–14), p \ 0.001] and a greater number of eGFR values \60 ml/min/1.73 m2 [medical units: 2 (1–4) vs. 1 (1, 2); surgery units: 4 (2–9) vs. 1 (1–3); ICU: 8 (3–14) vs. 2 (1–5), p \ 0.001]. Results did not differ among the different units (Table 2). Reporting of kidney dysfunction: main body vs. conclusion of HDS
combination were noted in 2.5 % of HDSs). Referral to a nephrologist (during hospitalization or planned referral after discharge) was indicated in 1.7 % of HDSs (Table 3). Influence of eGFR levels Among medical units, kidney dysfunction was reported in 10.3, 31.3, 68.8 and 89.5 % of HDSs for eGFR values 45–59, 30–44, 15–29 and \15 ml/min/1.73 m2, respectively. Similar results were observed for all units (Table 2). Using ROC curves, we found that the optimal cut-off for reporting kidney dysfunction in HDSs was 39 ml/min/ 1.73 m2 (AUC 0.81, sensitivity 0.69, specificity 0.80) in medical units; 40 ml/min/1.73 m2 (0.79, 0.71, and 0.76, respectively) in surgical units, and 36 ml/min/1.73 m2 (0.86, 0.81, and 0.83, respectively) in ICUs. Influence of duration of kidney dysfunction
In medical units, kidney dysfunction was reported in the conclusion of the HDS in only 11.4 % hospitalizations (medical units 9.8 %; surgical units 8.4 %; ICUs 27.5 %) (Table 2). AKI was noted in 7.0 % of HDSs, whereas CKD was noted in 13.1 % (both AKI and CKD in
123
When eGFR remained \60 ml/min/1.73 m2 at discharge, kidney dysfunction was reported in 35.8 % of HDSs in medical units, in 22.5 % in surgical units and in 62.2 % in ICUs (Table 4). The proportions were much lower when
49 (41–55) 49 (41–55) 32 (18–43) 33 (23–44) 50 (41–55)
5 (2–9)
3 (2–7)
31 (20–43)
Number of creatinine measurements, median (Q1–Q3)
Number eGFR values \60 ml/min/1.73 m2, median (Q1–Q3)
Lowest eGFR value, median (Q1–Q3)
2
10 (5–20) Duration of hospitalization, median (Q1–Q3) (days)
84.7 eGFR 0–14 ml/min/1.73 m2
Parameters associated with reporting of kidney dysfunction
65.3 eGFR 15–29 ml/min/1.73 m2
389 (11.4)
30.2 eGFR 30–44 ml/min/1.73 m2
Notification in conclusion of HDS, n (%)
9.7 eGFR 45–59 ml/min/1.73 m2
Reporting by eGFR value:
Yes, 866 (25.3)
HDS hospital discharge summary, eGFR estimated glomerular filtration rate (ml/min/1.73 m2), ICU intensive care unit, Q1–Q3 interquartile range
1 (1–3) 4 (2–9) 1 (1, 2) 2 (1–4)
Influence of age class
Kidney dysfunction was defined as a eGFR value \60 ml/min/1.73 m . All comparisons by units were significant (p \ 0.001)
2 (1–5) 7 (3–11)
1 (1–3)
kidney dysfunction was due to AKI (16.0, 17.1, and 37.7 % in medical, surgical units and ICUs, respectively).
23 (13–32)
19 (11–30) 7 (4–12) 9 (7–16) 4 (2–8)
2 (1–3) 3 (1–5)
7 (4–14) 6 (2–12)
2 (1–5)
212 (9.8)
89.5
50 (42–55)
103 (27.5) 74 (8.4)
9 (5–14)
91.4 61.7
2 (1–5)
79.8 38.5 68.8
8 (3–14)
43.9 22.3 31.3
10.3
Yes, 542 (25.0) No, 2,555 (74.7)
11 (7–20)
16.7 6.1
Yes, 181 (48.4) Yes, 143 (16.3) No, 1,627 (75.0)
No, 735 (83.7)
ICUs (n = 374)
Kidney dysfunction reported, n (%)
Table 2 Reporting of kidney dysfunction in HDSs
All units (n = 3,421)
Medical units (n = 2,169)
Surgical units (n = 878)
No, 193 (51.6)
14 (7–25)
J Nephrol
We divided patients into 3 age classes: 180 patients were under 50 years, 1,860 were 50–79 years, and 1,381 were over 80 years old. The reporting of kidney dysfunction according to age class and subdivided by level of eGFR and hospitalization unit is presented in Table 5. Low eGFR tended to be more frequently reported in patients aged \50 years. However, in patients with eGFR between 30 and 44 ml/min/1.73 m2, similar figures were observed regardless of age (Table 5). Overall, reporting was less than 30 % among patients aged \50 years (Table 5).
Discussion The results of our study indicate that kidney dysfunction frequently occurs during hospitalization, but is noted in the HDS in only one out of four patients. The rate of reporting of kidney dysfunction may vary by type of hospitalization unit (medical, surgical, or ICU), eGFR level, and duration of renal dysfunction (AKI vs. persisting renal dysfunction or CKD) but it remained low. The intention to convey information is reflected by the content of the HDS conclusion: the reporting of kidney dysfunction was present in only 11.4 % of HDS conclusions. Our study is one of the few assessing how kidney dysfunction is reported to primary-care physicians via the HDS. In 2007, a review of communication with the primarycare physicians at hospital discharge found that results of diagnostic tests were not mentioned in 33–63 % of HDSs, but the specific diagnostic tests missing were not detailed [1]. Our study instead gives complete information on the transmission of kidney dysfunction information at patient discharge via the HDS. The lack of knowledge of CKD has been investigated in different situations. In 2008, Rothberg et al. [13] showed that 46 % of CKD cases in an elderly population were recognized by physicians. In 2009, Gentile et al. [14] found that diagnostic codes were correctly used for only 19 % of hospitalized patients with CKD in an Italian hospital. The same year, Guessous et al. [15] estimated in a North American health system study that only 14.4 % (1,478/ 10,266) of patients with CKD were diagnosed. The lack of knowledge of CKD implies a delay in management of patients with increased hospitalizations and mortality [10]. Our study confirms the lack of knowledge regarding kidney dysfunction, and provides an analysis by type of hospitalization unit (medical, surgical and ICU). To improve the quality of care of patients with CKD, several institutions
123
J Nephrol Table 3 Kidney dysfunction reporting: specific information included in the HDS Units
All units (3,421)
Medical (2,169)
Surgical (878)
Intensive care (374)
AKI noted
239 (7.0)
145 (6.7)
87 (9.9)
137 (36.6)
CKD noted
449 (13.1)
336 (15.5)
44 (5.0)
69 (18.4)
Superimposed AKI on CKD
84 (2.5)
39 (1.8)
14 (1.6)
31 (8.3)
Planned referral to a nephrologist
59 (1.7)
45 (2.1)
3 (0.3)
11 (2.9)
Data are number (%) HDS hospital discharge summary, AKI acute kidney injury, CKD chronic kidney disease
Table 4 Reporting of kidney dysfunction in the main body and conclusion of the HDS according to the type of kidney dysfunction (persistent kidney dysfunction vs. acute kidney injury) Units
All units (n = 2,328)
Medical (n = 1,314)
Surgical (n = 651)
Intensive care (n = 363)
Kidney dysfunction reported
Yes
p value
Yes
p value
Yes
p value
Yes
p value
514 (35.6)
\0.001
334 (35.8)
\0.001
78 (22.5)
0.087
102 (62.2)
\0.001
In the main body of HDS, n (%) Persistent kidney dysfunction at discharge Acute kidney injury In the conclusion of HDS, n (%)
188 (21.2)
Persistent kidney dysfunction at discharge
236 (16.4)
Acute kidney injury
61 (16.0) 0.854
130 (13.9)
88 (9.9)
52 (17.1) 0.673
22 (5.8)
36 (10.4) 36 (11.8)
75 (37.7) 0.051
70 (42.7)
\0.001
30 (15.1)
Of note, hospitalizations were selected when serum creatinine was measured more than once HDS hospital discharge summary
Table 5 Reporting of kidney dysfunction in the HDS according to age class
Kidney dysfunction reported, n (%) eGFR (ml/min/1.73 m2)
30–44 (n = 935)
15–29 (n = 406)
28/99 (28.3)
15/41 (36.6)
16/21 (76.2)
Medical units
16/46 (34.8)
10/14 (71.4)
4/7 (57.1)
3/3 (100.0)
Surgical units
3/27 (11.1)
2/4 (50.0)
2/4 (50.0)
5/7 (71.4)
Intensive care units
9/26 (34.6)
3/3 (100.0)
10/10 (100.0)
7/9 (77.8)
\50 years (n = 180) All units
45–59 (n = 1,918)
0–14 (n = 162)
15/19 (78.9)
50–79 years (n = 1,860) All units
177/1,153 (15.4)
164/444 (36.9)
122/171 (71.3)
78/92 (84.8)
Medical units
127/724 (17.5)
106/275 (38.5)
72/96 (75.0)
21/24 (87.5)
Surgical units
24/333 (7.2)
28/110 (25.5)
12/31 (38.7)
18/29 (62.1)
30/59 (50.8)
38/44 (86.4)
39/39 (100.0)
Intensive care units
26/96 (27.1)
[80 years (n = 1,381)
HDS hospital discharge summary, eGFR estimated glomerular filtration rate
All units
147/666 (22.1)
172/450 (38.2)
142/214 (66.4)
46/51 (90.2)
Medical units
132/478 (27.6)
140/320 (43.8)
102/141 (72.3)
28/30 (93.3)
Surgical units
11/35 (31.4)
24/114 (21.1)
18/43 (41.9)
8/11 (72.7)
4/28 (14.3)
8/16 (50.0)
22/30 (73.3)
10/10 (100.0)
Intensive care units
such as the US National Kidney Foundation recommended the systematic estimation of GFR in medical laboratories for each creatinine serum measurement [16]. In 2010, Hemmelgarn et al. [17] estimating the impact of systematic estimation of GFR showed an increase in nephrology consultations; however, no meaningful consequences were noted for patients. In 2012, Akbari et al. [18] showed an
123
increase in nephrology consultations after the systematic estimation of GFR but also an increase in inappropriate consultations. Patients with kidney dysfunction during hospitalization were seldom diagnosed, with disparity among units. Awareness of CKD has improved with the systematic estimation of GFR by laboratories. However, specific training
J Nephrol
of medical professionals to increase the rate of detection and adequate management of kidney dysfunction is also important. A study performed in 2004 by Akbari et al. [19] analyzed an educational program of primary-care providers and showed an improvement in detection of CKD. Overall, the lack of knowledge of CKD involves a delay in management of patients with increased hospitalizations and mortality, and missed opportunities to initiate appropriate treatment strategies. The training of hospital physicians to detect kidney disease and convey this diagnostic information to primary-care physicians seems an important step. Another step is the adequate training of primary-care physicians to manage patients with CKD within the framework of an educational public health program. We assessed eGFR during hospitalization but we had no data to define AKI or CKD with precision in the present study. Moreover, the severity indicated by the level of eGFR differs according to the patient’s age and comorbidities. The aim of our study was to describe the reporting of kidney dysfunction, not to ensure that the diagnosis of AKI or CKD was accurate, nor to assess the severity of kidney dysfunction. To estimate renal function, we used the MDRD formula rather than the calculation of creatinine clearance by the Cockcroft–Gault formula because bodyweight data was missing in most of the hospitalizations. Of note, GFR estimated by the MDRD formula is better in subjects in stable conditions, and the estimate may be biased in acutely ill patients; however, there is presently no adequate GFR formula in this situation [20]. We analyzed only hospitalizations with creatinine serum measurements. Given that more that 90 % of patients had a serum creatinine measurement in medical units and ICUs, the rate of kidney dysfunction we observed can be considered as the true prevalence in these units. Our study is a monocentric study performed in a university hospital: is it unknown whether communication regarding kidney dysfunction may be different in non-university hospitals. We did not analyze management of kidney dysfunction during hospitalizations: although this aspect is important, it was beyond the scope of our study. Importantly, systematic estimation of eGFR was not performed in our hospital, and thus it was not possible to assess whether systematic eGFR reporting could modify these findings. Whether automatic reporting of eGFR would modify the HDS reporting is presently unknown but warrants investigation. Kidney dysfunction is frequent but its reporting in the HDS is often missing, suggesting that this information is usually not considered important to disclose to primarycare physicians. The absence of this information in the HDS conclusions should be analyzed in other settings. The consequences of our observations should be studied in terms of continuity of patient care after discharge and late referral to nephrologists.
Conflict of interest of interest. Ethical standard required.
The authors declare that they have no conflict
For this type of study formal consent is not
References 1. Kripalani SLF (2007) Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA 297(8):831–841 2. Wang HE, Muntner P, Chertow GM, Warnock DG (2012) Acute kidney injury and mortality in hospitalized patients. Am J Nephrol 35(4):349–355 3. Coca SG, Singanamala S, Parikh CR (2012) Chronic kidney disease after acute kidney injury: a systematic review and metaanalysis. Kidney Int 81(5):442–448 4. James MT, Hemmelgarn BR, Tonelli M (2010) Early recognition and prevention of chronic kidney disease. Lancet 375(9722):1296–1309 5. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu C (2004) Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 351(13):1296–1305 6. Stevens LA, Fares G, Fleming J, Martin D, Murthy K, Stark PC, Uhlig K, Van Lente F, Levey AS (2005) Low rates of testing and diagnostic codes usage in a commercial clinical laboratory: evidence for lack of physician awareness of chronic kidney disease. J Am Soc Nephrol 16(8):2439–2448 7. Jungers P (2002) Late referral: loss of chance for the patient, loss of money for society. Nephrol Dial Transplant 17(3):371–375 8. Chan MR, Dall AT, Fletcher KE, Lu N, Trivedi H (2007) Outcomes in patients with chronic kidney disease referred late to nephrologists: a meta-analysis. Am J Med 120(12):1063–1070 9. Smart NA, Titus TT (2011) Outcomes of early versus late nephrology referral in chronic kidney disease: a systematic review. Am J Med 124(11):1073–1080.e2 10. Levey AS, Coresh J (2012) Chronic kidney disease. Lancet 379(9811):165–180 11. Kim DH, Kim M, Kim H, Kim YL, Kang CW, Kim NH, Kim YS, Lee JP (2013) Early referral to a nephrologist improved patient survival: prospective cohort study for end-stage renal disease in Korea. PLoS ONE 8(1):e55323 12. Palevsky PM, Liu KD, Brophy PD, Chawla LS, Parikh CR, Thakar CV, Tolwani AJ, Waikar SS, Weisbord SD (2013) KDOQI US commentary on the 2012 KDIGO clinical practice guideline for acute kidney injury. Am J Kidney Dis 61(5):649–672 13. Rothberg MB, Kehoe ED, Courtemanche AL, Kenosi T, Pekow PS, Brennan MJ, Mulhern JG, Braden GL (2008) Recognition and management of chronic kidney disease in an elderly ambulatory population. J Gen Intern Med 23(8):1125–1130 14. Gentile G, Postorino M, Mooring RD, De Angelis L, Manfreda VM, Ruffini F, Pioppo M, Quintaliani G (2009) Estimated GFR reporting is not sufficient to allow detection of chronic kidney disease in an Italian regional hospital. BMC Nephrol 10:24 15. Guessous I, McClellan W, Vupputuri S, Wasse H (2009) Low documentation of chronic kidney disease among high-risk patients in a managed care population: a retrospective cohort study. BMC Nephrol 10:25 16. National Kidney Foundation (2002) K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 39(Suppl 1):S1–S266 17. Hemmelgarn BR, Zhang J (2010) Nephrology visits and health care resource use before and after reporting estimated glomerular filtration rate. JAMA 303(12):1151–1158
123
J Nephrol 18. Akbari A, Grimshaw J, Stacey D, Hogg W, Ramsay T, ChengFitzpatrick M, Magner P, Bell R, Karpinski J (2012) Change in appropriate referrals to nephrologists after the introduction of automatic reporting of the estimated glomerular filtration rate. CMAJ 184(5):E269–E276 19. Akbari ASP (2004) Detection of chronic kidney disease with laboratory reporting of estimated glomerular filtration rate and an educational program. Arch Intern Med 164(16):1788–1792
123
20. Poggio ED, Nef PC, Wang X, Greene T, Van Lente F, Dennis VW, Hall PM (2005) Performance of the Cockcroft–Gault and modification of diet in renal disease equations in estimating GFR in ill hospitalized patients. Am J Kidney Dis 46(2):242–252
