This project was supported, in part, by a Dennis W. Jahnigen Career Development Scholar Award (JMC) from the American Geriatrics Society, John A. Hartford Foundation, and Atlantic Philanthropies.
Original Research Contribution
Age, Nursing Home Residence, and Presentation of Urinary Tract Infection in U.S. Emergency Departments, 2001–2008
Article first published online: 15 OCT 2012
© 2012 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 19, Issue 10, pages 1173–1180, October 2012
How to Cite
Caterino, J. M., Ting, S. A., Sisbarro, S. G., Espinola, J. A. and Camargo, C. A. (2012), Age, Nursing Home Residence, and Presentation of Urinary Tract Infection in U.S. Emergency Departments, 2001–2008. Academic Emergency Medicine, 19: 1173–1180. doi: 10.1111/j.1553-2712.2012.01452.x
The authors have no potential conflicts of interest to disclose.
Supervising Editor: Lowell Gerson, PhD.
- Issue published online: 15 OCT 2012
- Article first published online: 15 OCT 2012
- Received March 27, 2012; revisions received April 13 and May 7, 2012; accepted May 10, 2012.
ACADEMIC EMERGENCY MEDICINE 2012; 19:1173–1180 © 2012 by the Society for Academic Emergency Medicine
Objectives: Current outpatient diagnostic algorithms for urinary tract infection (UTI) in older adults require the presence of classic signs and symptoms of UTI, such as fever and genitourinary symptoms. However, older adults with UTI may present with atypical signs and symptoms. The objective was to identify the associations of age and nursing home status with the clinical presentation of emergency department (ED) patients diagnosed with UTI.
Methods: This was a retrospective, cross-sectional analysis of the 2001–2008 National Hospital Ambulatory Medical Care Survey (NHAMCS), ED component. Participants were adult ED patients diagnosed with UTI. Outcome variables were presence of fever, altered mental status, and urinary tract symptoms. Multivariable logistic regression models were constructed for each outcome. Age and nursing home status were the independent variables of interest. Age was divided into adults 18 to 64 years, older adults 65 to 84 years, and oldest adults 85 years of age and older.
Results: There were 25.4 million ED visits in which UTI was diagnosed from 2001 through 2008, including 5.0 million in older adults and 2.2 million in the oldest adults. Fever was present in 13% of adults, 21% of older adults, and 19% of the oldest adults. Altered mental status was present in 1% of adults, 7% of older adults, and 13% of the oldest adults. Urinary tract symptoms were identified in 32% of adults, 24% of older adults, and 17% of the oldest adults. In multivariable analysis, altered mental status was more common in older adults (odds ratio [OR] = 1.94) and in the oldest adults (OR = 2.49). Urinary tract symptoms were less common in older adults (OR = 0.60) and the oldest adults (OR = 0.48). Nursing home residence was associated with increased fever (OR = 1.63) and altered mental status (OR = 4.79) and with decreased urinary tract symptoms (OR = 0.35).
Conclusions: Fever and urinary tract symptoms are absent in a large proportion of adults over 65 years of age diagnosed with UTI in the ED. Age over 65 years and nursing home residence are associated with increased presence of altered mental status and with lack of urinary tract symptoms. Nursing home residence is also associated with increased presence of fever. Emergency physicians (EPs) continue to diagnose UTI in patients without classic symptoms. Diagnostic criteria for UTI among adults 65 years and over specifically designed for use in the acute care setting should be developed and validated to prevent both inappropriate overdiagnosis and underdiagnosis of UTI.
Objetivos: Los actuales algoritmos diagnósticos de la infección del tracto urinario (ITU) en los adultos mayores ambulatorios requieren la presencia de los síntomas y signos clásicos de ITU, como la fiebre y los síntomas genitourinarios. Sin embargo, los adultos mayores con ITU pueden presentase con síntomas y signos atípicos. El objetivo fue identificar la asociacion de la edad y la institucionalización con la presentación clínica de los pacientes diagnosticados con ITU en el servicio de urgencias (SU).
Métodos: Estudio transversal retrospectivo del área del SU del National Hospital Ambulatory Medical Care Survey entre 2001 y 2008. Los participantes fueron pacientes adultos del SU diagnosticados de ITU. Las variables resultado fueron la presencia de fiebre, la alteración del estado mental y los síntomas del tracto urinario. Se utilizó modelos de regresión logística para cada resultado. La edad y la institucionalización fueron las variables independientes de interés. La edad se dividió en adultos de 18 a 64 años, ancianos de 65 a 84 años, y muy ancianos de 85 años o más.
Resultados: Hubo 25,4 millones de visitas a SU con el diagnóstico de ITU desde 2001 a 2008, que incluyó 5.0 millones de ancianos y 2,2 millones muy ancianos. La fiebre se presentó en el 13% de los adultos, el 21% de los ancianos, y el 19% de los muy ancianos. La alteración del estado mental se presentó en el 1% de los adultos, el 7% de los ancianos, y el 13% de los muy ancianos. Los síntomas urinarios se identificaron en el 32% de los adultos, el 24% de los ancianos, y el 17% de los muy ancianos. En el análisis multivariable, la alteración del estado mental fue más común en los ancianos (odds ratio [OR] 1,94) y en los muy ancianos (OR 2,49). Los síntomas urinarios fueron menos comunes en los ancianos (OR 0,60) y en los muy ancianos (OR 0,48). La institucionalización se asoció con un incremento de la fiebre (OR 1,63) y del estado mental alterado (OR 4,79), y con una disminución de los síntomas del tracto urinario (OR 0,35).
Conclusiones: La fiebre y los síntomas del tracto urinario están ausentes en una proporción alta de adultos mayores de 65 años diagnosticados de ITU en el SU. La edad mayor de 65 años y la institucionalización se asociaron con un incremento de la presencia de alteración del estado mental y con una ausencia de los síntomas del tracto urinario. La institucionalización también se asoció con un incremento de la presencia de fiebre. Los urgenciólogos continúan diagnosticando ITU en los pacientes sin síntomas clásicos. Deberían desarrollarse criterios diagnósticos de ITU específicamente diseñados para el uso en la atención aguda de los adultos de 65 o más años para prevenir tanto el inapropiado sobrediagnóstico como el infradiagnóstico de ITU.
Each year, patients aged 65 years and over account for approximately 500,000 visits for urinary tract infection (UTI) to U.S. emergency departments (EDs).1 These infections include both cystitis and pyelonephritis. Adults 65 years and over may be less likely to present with “classic” symptoms of UTI, such as fever and dysuria, than younger adults.2–5 Additionally, adults 65 years and over may be more likely to present with nonspecific symptoms such as altered mental status. This complicates care for these patients, as there are currently no widely accepted criteria for diagnosis of UTI in adults 65 years and over in the ED setting. Studies have found that urine dipsticks are often inaccurate in this population, culture results are generally not available, and the test characteristics of both genitourinary and atypical symptoms are unclear.2,6,7
Diagnostic criteria for UTI have been developed in nursing home settings to guide physician decision-making in initiating antimicrobial therapy.8 The focus of these guidelines is to “lead to a substantial reduction in inappropriate use of antibiotics” in the long-term care setting.8 Such a conservative approach may not be appropriate in acutely ill older adults in the ED. For example, the criteria are insensitive in identifying bacteremic UTI in the inpatient setting.4 These guidelines also require the presence of specific UTI symptoms, but it is not clear how often these are present in the ED setting.2 Therefore, understanding the presenting symptoms among patients diagnosed with UTI in the ED will provide information on the need to further develop guidelines for the acute care setting.
Studies on the presenting symptoms of UTI in adults 65 and over have largely been conducted either in community-dwelling or nursing home patients.5,9–11 Patients in acute care settings are likely very different from either of these populations, as they may represent a more severely ill subset, and they represent a combination of patients arriving from both the community and the nursing home.12 Presentation in nursing home patients can be very different from that in the community.9 Additionally, the oldest adults (those 85 years and over) may have different presentations for UTI than older adults 65 to 84 years old.13
We sought to better understand the role that advancing age and nursing home residence may have on the clinical presentation of patients diagnosed with UTI among adults 65 years and over presenting to the ED. We studied three symptoms: fever, altered mental status, and urinary tract symptoms. We hypothesized that adults 65 years and over would be less likely to present with fever and urinary tract symptoms, and more likely to have altered mental status, than younger adults. We also hypothesized that these patterns would be more pronounced in nursing home residents.
We conducted a cross-sectional analysis of the ED component of the 2001 through 2008 National Hospital Ambulatory Medical Care Survey (NHAMCS) database. NHAMCS was designed by the National Center for Health Statistics of the Centers for Disease Control and Prevention and is a national probability survey conducted for both hospital outpatient and ED visits. The hospital’s institutional review board determined that this study did not require review since the use of deidentified registry data does not constitute human subjects research as defined in 45 CFR 46:102(f). This work was supported, in part, by a Dennis W. Jahnigen Career Development Award sponsored by the American Geriatrics Society, John A. Hartford Foundation, and Atlantic Philanthropies. Funding agencies had no role in the design, execution, analysis and interpretation of the data, or writing of the study.
Study Setting and Population
The NHAMCS is a four-stage probability sample survey gathering data from noninstitutional general and short-stay hospitals, excluding federal, military, and Veterans Administration hospitals.14 Data are collected annually by trained hospital staff during a randomly assigned 4-week data period for each of the sampled hospitals.15 Quality control measures included checks of value ranges, and adjudication for ambiguous or illegible responses. The nonresponse rate for most items was <5%, and error rates were <2% for items requiring medical coding.14 National estimates were obtained through use of a multistage estimation procedure and patient visit weights. A more detailed description of the NHAMCS data collection and estimation procedures is available online.14,16
Our study population included all ED visits by patients age ≥18 years in the 2001–2008 NHAMCS database with an International Classification of Disease, Ninth Revision–Clinical Modification (ICD-9 CM) code in any of the three diagnosis fields for UTI including cystitis and pyelonephritis (590, 595.0, 595.89, 595.9, or 599.0)
The three outcomes studied were presence of fever, altered mental status, and urinary tract symptoms. NHAMCS includes ICD-9CM diagnosis codes (up to three codes), certain specific vital signs and physical exam fields, and reason for visit codes (up to three codes). Reasons for visit are coded using the reason for visit classification and coding manual.17 Presence of fever was defined by an ICD-9 diagnosis code 780.6; documented temperature (field for variable TEMPF) ≥100.4°F; or any reason for visit codes 1005.0, 1010.0, or 1012.0. Altered mental status was defined as an ICD-9 code 780.97; documentation of disorientation (field for variable ORIENTED = no); or presence of reason for visit codes 5840.0, 5841.0, or 5842.0. Presence of urinary tract symptoms was determined through the presence of any of the following reason for visit codes: 1640.0–1640.3, 1645.0, 1645.1, 1650.0, 1655.0, 1655.1, 1660.1–1660.4, 1665.0–1665.2, 1670.0, 1670.1, 1670.2, 1675.0, or 1710.0–1710.2.
The primary independent variables of interest were age and nursing home status. Patients were stratified into adults (aged 18 to 64 years), older adults (aged 65 to 84 years), and oldest adults (aged ≥85 years). Residence was considered a nursing home if coded as residing in a nursing home or other institution. Additional potential independent variables included sex, race (white, black or African American, or other), Hispanic ethnicity, and source of payment (private, public, self-pay, no charge/charity care, other/unknown). Hospital-level data included metropolitan statistical area (MSA) status and U.S. region (Northeast, Midwest, South, and West) as defined by the U.S. Bureau of the Census. Disposition was defined as dead on arrival/died in ED, transferred to other facility, admitted to hospital, retained for 23-hour observation, or discharged.
To account for the complex four-stage sampling frame, all analyses were performed using survey design variables and appropriate survey commands in STATA 11.2 (StataCorp, College Station, TX). We determined point estimates and 95% confidence intervals (CIs) for visit characteristics overall and stratified by age group (adults 18 to 64 years, older adults 65 to 84 years, and oldest adults ≥85 years). We also obtained estimates for the entire population ≥65 years stratified by nursing home status.
We constructed a separate multivariable model for each of the three outcome variables. We first tested potential independent variables in a series of weighted unadjusted logistic regression models. All variables with a univariate p < 0.3 were retained in standard multivariable logistic regression models. No stepwise regression was used. Age group and nursing home status were included in each multivariable model regardless of statistical significance, as they were the primary independent variables of interest. All models were tested for multicolinearity, interactions, and for goodness of fit using the Hosmer-Lemeshow test. We did not test for extreme outliers. As we were constructing three discrete multivariable models, we used a p-value of 0.16 as an indicator of significance based on the Bonferroni correction for multiple comparisons.
From 2001 through 2008, we identified 7,730 NHAMCS records meeting study entry criteria. Accounting for their weights, they represented approximately 25.4 million ED visits for UTI. This was 3.6% (95% CI = 3.5% to 3.8%) of total ED visits during the 8-year period. A total of 18.2 million visits were by adults, 5.0 million by older adults, and 2.2 million by the oldest adults. The percentage of visits attributable to UTI increased with age, from 3.2% (95% CI = 3.1% to 3.4%) of all ED visits in adults, to 4.7% (95% CI = 4.4% to 4.9%) in older adults, and to 7.6% (95% CI = 6.9% to 8.3%) in the oldest adults.
Characteristics of the study population, overall and stratified by age, are shown in Tables 1A and 1B. Overall, 15% of patients had fever at triage. This number increased from 13% in adults to 21% in older adults and 19% in the oldest adults. Altered mental status was noted in 3% of study patients overall, but in only 0.8% of adults. By contrast, 7% of older adults and 13% of the oldest adults had altered mental status. Urinary tract symptoms were noted in 25% of the study population and were more common in adults (32%) than in older adults (24%) and the oldest adults (17%). Patients age 65 and older were more likely to be admitted.
|Characteristics||Adults 18–64 years||Older adults 65–84 years|
|No. of Visits (Thousands)||95% CI (Thousands)||% of Study Population Age 18–64 yr||No. of Visits (Thousands)||95% CI (Thousands)||% of Study Population Age 65–84 yr|
|Overall number of visits||18,200||15,900–20,400||(100)||5,015||4,354–5,676||(100)|
|Black or African American||4,775||3,993–5,557||26.2||661||509–813||13.2|
|Altered mental status||143||92–193||0.8||367||255–478||7.3|
|Urinary tract symptoms||5,885||5,131–6,640||32.3||1,200||1,005–1,394||23.9|
|Not in MSA||3,028||1,772–4,284||16.6||979||578–1,381||19.5|
|Source of payment|
|DOA/died in ED||3.7||nc||0.0||0.6||nc||0.0|
|Transfer to other facility||130||79–180||0.7||103||59–146||2.1|
|Admit to hospital/ICU||1,381||1,143–1,618||7.6||1,931||1,635–2,228||38.5|
|Admit for 23-hour observation||140||84–196||0.8||75||nc||1.5|
|Characteristics||Oldest Adults 85 Years of Age and Older||Entire Population 18 Years and Over|
|No. of Visits (Thousands)||95% CI (Thousands)||% of Study Population Age ≥85 yr||No. of Visits (Thousands)||95% CI (Thousands)||% of Study Population|
|Overall number of visits||2,203||1,880–2,526||(100)||25,400||22,300–28,500||(100)|
|Black or African American||240||165–314||10.9||5,676||4,749–6,603||22.3|
|Altered mental status||293||211–375||13.3||803||638–968||3.2|
|Urinary tract symptoms||370||273–468||16.8||7,455||6,541–8,369||29.4|
|Not in MSA||382||218–546||17.3||4,389||2,643–6,135||17.3|
|Source of payment|
|DOA/died in ED||6.4||nc||0.3||11||nc||0.0|
|Transfer to other facility||12||nc||0.5||245||170–319||1.0|
|Admit to hospital/ICU||1,201||1,004–1,399||54.5||4,514||3,904–5,123||17.8|
|Admit for 23 hour observation||40||nc||1.8||255||171–339||1.0|
We examined the effect of nursing home residence on presentation in the population ≥65 years of age. Among all patients age ≥65 years, nursing home patients were slightly more likely to have fever (26%, 95% CI = 24% to 28%) than community dwellers (19%, 95% CI = 18% to 20%). Nursing home patients were also more likely to have altered mental status (23%, 95% CI = 21% to 25%) than community dwellers (5%, 95% CI = 4% to 6%). However, urinary tract symptoms were less common in nursing home patients (9%, 95% CI = 4% to 16%) than community dwellers (26%, 95% CI = 25% to 28%). Nursing home patients were also more likely to be admitted to the hospital (60%, 95% CI = 58% to 62%) than community dwellers (37%, 95% CI = 36% to 38%).
Results of the unadjusted logistic regression analyses conducted in the entire study population are shown in Table 2 (complete results available in Data Supplement S1, available as supporting information in the online version of this paper). In these unadjusted analyses, both age group and nursing home status were significantly related to all outcomes.
|Characteristics||Fever||Altered Mental Status||Urinary Tract Symptoms|
|OR||95% CI||p-value||OR||95% CI||p-value||OR||95% CI||p-value|
|Age, yr (overall)|
|65–84 (older adults)||0.95||0.78–1.15||0.60||1.89||1.19–3.01||0.007||0.67||0.53–0.83||<0.001|
|>85 (oldest adults)||0.76||0.59–0.97||0.03||2.90||1.85–4.53||<0.001||0.47||0.35–0.64||<0.001|
|Nursing home resident|
Results of the multivariable logistic regression models are reported for each outcome in Table 3 (complete results available in Data Supplement S2, available as supporting information in the online version of this paper). In the model for fever, nursing home residents were more likely than community dwellers to present to the ED with a fever (adjusted odds ratio [OR] = 1.63, 95% CI = 1.18 to 2.25). Age group was not significantly associated with fever.
|Characteristics||Fever||Altered Mental Status||Urinary Tract Symptoms|
|Adjusted OR||95% CI||p-value||Adjusted OR||95% CI||p-value||Adjusted OR||95% CI||p-value|
|65–84 (older adults)||0.87||0.64–1.18||0.37||1.94||0.99–3.82||0.06||0.60||0.39–0.92||0.02|
|>85 (oldest adults)||0.74||0.52–1.03||0.08||2.49||1.25–4.95||0.009||0.48||0.31–0.75||0.001|
|Nursing home resident|
Both increasing age group and nursing home status were associated with increased odds of altered mental status. Compared to adults 18 to 64 years, the oldest adults were more likely to present to the ED with altered mental status. Older adults demonstrated a trend toward an increase in altered mental status, but this was borderline significant (adjusted OR = 1.94, 95% CI = 0.99 to 3.82; p = 0.06). In addition, nursing home residents had more documented ED visits with altered mental status than community dwellers.
In the model for urinary tract symptoms, both increasing age group and nursing home status were associated with decreased odds of symptoms. When compared to adults, both older adults and the oldest adults were significantly less likely to present with urinary tract symptoms. Nursing home residents presented less frequently with urinary tract symptoms than community dwellers. No multicolinearity or significant interactions were identified in any of the multivariate models. Fit was adequate in each model according to the Hosmer-Lemeshow test, with all p-values >0.05.
In a large nationally representative database, we found that presentation of patients diagnosed with UTI in the ED differs by both age and nursing home status. Nursing home residence was associated with increased presence of fever. Advanced age and nursing home residence were also associated with a significantly greater likelihood of altered mental status and lower likelihood of urinary tract symptoms.
Our results represent several advances over the current understanding of ED diagnosis of UTI in older adults. First, by providing national estimates, we extend the findings of single academic center studies.2–4,6,12 Second, these studies generally examine older adults only and do not compare them to younger adults. We provide data on three age groups, demonstrating the clear difference and the magnitude of the difference between younger and older adults in the ED setting. Third, studies have been done in the inpatient setting, which support our findings, but it is not immediately clear that inpatient data can be extrapolated to the entire ED population.3,4 Finally, inclusion criteria in these studies have not represented the full spectrum of ED patients diagnosed with UTI. For example, subjects have variously been limited to only those with positive cultures,2–4,12 only those with UTI symptoms,12 or only those without UTI symptoms.6 By limiting study inclusion in this way, a complete picture of emergency physician (EP) diagnosis of UTI in older adults is elusive. By including all ED patients diagnosed with UTI, we identify EP diagnostic behavior on a national level. This provides information on the overall real-world diagnosis of UTI in U.S. EDs, which has not previously been reported.
Our results from a multiyear national sample of all patients diagnosed with UTI are consistent with previous single-center studies of age and UTI in which positive urine cultures were required for inclusion. Ginde et al.2 studied ED older adults with UTI and found similar rates of urinary symptoms (26%) and fever (17%) among patients with positive urine cultures. Woodford and George3 found the same patterns of symptoms among inpatients with bacteremic UTI. By providing results of a national sample, we provide assurance that these prior single-center findings regarding effect of age on presentation of UTI are generalizable to the acute care ED setting. We also confirm that, regardless of ultimate culture results, EPs are currently acting on their belief that older adults with UTI need not have genitourinary symptoms but may have altered mental status in making their diagnosis.
We also examined the relationship between nursing home residence and presenting symptoms. Nursing home residence magnified the effects of age, as nursing home patients were independently more likely to have fever and altered mental status, but less likely to have genitourinary symptoms. The magnitude of the effect of nursing home residence was greater than that of age. A nursing home resident had 4.79 times the odds of having altered mental status and only 0.35 times the odds of having urinary tract symptoms as a nonresident. Both of these adjusted ORs are greater in magnitude than those for age 65 to 84 years and ≥85 years. Our findings are consistent with prior single-center studies in the nursing home, and they confirm that these patterns also apply to nursing home patients sent to the ED.5,9 Our findings point out the need to be particularly vigilant of the absence of classic symptoms in nursing home patients presenting to the ED.
Perhaps our most important finding comes from identifying the need for development of UTI diagnostic criteria for older adults in the ED and acute care setting. In our study, a large number of adults over 65 years of age diagnosed as having UTI in the ED did not have classic symptoms such as fever or urinary tract symptoms. The absence of these classic symptoms could raise questions regarding the accuracy of ED diagnosis, particularly in the absence of urine culture results, which are not available in the ED.
Physicians often struggle with the empiric diagnosis of UTI both due to inadequate diagnostics and inadequate definitions. The diagnosis is complicated by the frequent presence of asymptomatic bacteriuria in elders.6,18,19 Diagnostically, urine dipsticks have shown to be poorly predictive of positive urine cultures in multiple settings.6,7,20 Additionally, a positive culture itself may represent asymptomatic bacteriuria rather than acute infection, and culture results are rarely available in the ED.18,19
Definitions for UTI in the acute care setting are also lacking. There are no currently accepted criteria for empiric diagnosis of UTI in older patients in the acute care setting. In the nursing home setting, authors have generally been concerned with distinguishing acute UTI from asymptomatic bacteriuria, a phenomenon also seen in ED patients.6,21 Driven by a desire to prevent overuse of antibiotics, minimum criteria (the Loeb Criteria) that rely largely on the presence of classic symptoms have been developed to indicate the need for empiric antibiotic therapy in nursing home patients with suspected UTI.8 Given the large proportions of patients without classic symptoms in our national sample, it is clear that the Loeb criteria are not currently followed in U.S. EDs.
However, this does not necessarily indicate the presence of inappropriate diagnoses or antibiotic prescribing. The applicability of the Loeb Criteria to acute care settings is unclear. In inpatients ≥75 years of age with proven UTI, Woodford and George3 found that only 49% had urinary tract symptoms, 47% had new or worsened confusion, and 40% had fever. Similar results have been found in inpatients with bacteremic UTI.4 Bacteremic UTI, isolation of the same pathogen from urine and blood, provides the strongest confirmation that an acute infection is truly present. The Loeb Criteria would have missed 43% of cases of bacteremic UTI.4 Our study in the ED setting is consistent with these prior inpatient findings in which a large number of older adults lack fever and dysuria, but have altered mental status.
Ultimately, our results reinforce the need for further study, clarification, and consensus in diagnosing acute UTI among adults over 65 years in the acute care setting and in differentiating it from asymptomatic bacteriuria. Given the lack of a current diagnostic criterion standard, these studies will likely require expert physician review of cases to determine the presence of acute infection. Identification of patient and test characteristics in a prospective manner using an established criterion standard for acute infection would allow identification of factors that support the diagnosis of acute UTI. These factors could ultimately be incorporated into diagnostic criteria or clinical decision rules, which would also require validation. EPs clearly diagnose UTI without evidence of classic symptoms in current clinical practice, although this is likely appropriate in many cases given the above noted inpatient studies. Ensuring accurate diagnoses is important for two reasons. First, inappropriate overdiagnosis of UTI may cause EPs to miss other etiologies for the patient’s symptoms and result in unnecessary exposure to antibiotics.8 Second, underdiagnosis could result in failure to provide timely, appropriate, empiric antibiotic therapy.13 Ideally, diagnostic criteria specific to the acute care setting could be derived and validated for diagnosis of UTI in older adults.
The NHAMCS database depends on a chart review methodology that may result in inaccurate or incomplete documentation of symptoms. We addressed this by including potential documentation of symptoms from multiple sources in the data set. However, we recognize that this may not fully capture the full spectrum of patient symptoms. Also, there is the possibility of diagnostic bias, as the diagnosis of UTI was based on final diagnoses by the EP, which is made without benefit of urine culture results.
As a result, we are unable to determine the diagnostic accuracy of the EPs. As noted above, even in the presence of culture results, a true criterion standard, may be elusive for UTI in older adults in the acute care setting. By including all those diagnosed in the ED, we have information on current ED clinical practice, but we cannot confirm the accuracy of that practice. Additionally, as we only enrolled patients diagnosed with UTI, we were unable to determine how many patients were missed by the ED (sensitivity of ED diagnosis). This may have biased results away from characteristics found in patients more difficult to diagnose, for example, in those without classic symptoms. Further studies are required to consider both the characteristics of patients diagnosed in the ED and their ultimate outcomes, to develop a true criterion standard.
Due to limitations in the data set, we were unable to assess comorbidities, recent urologic procedures, recent or current antibiotic use, or the presence of indwelling catheters and the effect these may have had on diagnosis of UTI. The presence of these factors may have certainly affected EP diagnostic thresholds, likely increasing the likelihood of UTI diagnosis.
Our measure of mental status included a diagnosis code for altered mental status, a reason for visit code of altered mental status, and a documentation of disorientation. Particularly for the latter, we are unable to confirm whether altered mental status was acute or chronic (e.g., delirium or dementia). For acute UTI, we are most interested in the former, i.e., acute changes in mental status. The inclusion of chronic conditions such as dementia may have artificially increased the presence of altered mental status in older adults. We attempted to limit this by including measures likely to represent true acute altered mental status. For example, the reason for visit codes should ideally be documented only when altered mental status was one of the reasons for the visit, not just a chronic condition. Finally, nursing home patients with fever may be more likely to be sent to the ED. As a result, these conclusions should not be extrapolated to the nursing home setting itself.
Fever and urinary tract symptoms are absent in a large proportion of adults 65 years and over diagnosed with UTI in the ED. Adults 65 years and over are less likely to have urinary tract symptoms than younger adults. Among the oldest old, altered mental status becomes more likely. Nursing home residence is associated with decreases in urinary tract symptoms, but increases in fever and altered mental status. EPs continue to diagnose UTI in patients without classic symptoms, but this study’s findings are similar to studies using confirmed diagnoses made in the inpatient setting. Diagnostic criteria for UTI among adults 65 years and over specifically designed for use in the acute care setting should be developed and validated to prevent both inappropriate overdiagnosis and underdiagnosis of UTIs in adults.
- 6Can urine cultures and reagent test strips be used to diagnose urinary tract infection in elderly emergency department patients without focal urinary symptoms? CJEM. 2007; 9:87–92., , .
- 14Centers for Disease Control and Prevention, National Center for Health Statistics. Ambulatory Health Care Data: About the Ambulatory Health Care Surveys. Available at: http://www.cdc.gov/nchs/ahcd/about_ahcd.htm. Accessed Jul 23, 2012.
- 15Centers for Disease Control and Prevention. National Center for Health Statistics. Ambulatory Health Care Data: Scope and Sample Design. Available at: http://www.cdc.gov/nchs/ahcd/ahcd_scope.htm#nhamcs_scope. Accessed Jul 23, 2012.
- 16Centers for Disease Control and Prevention. National Center for Health Statistics. Ambulatory Health Care Data: NHAMCS Data Collection and Processing. Available at: http://www.cdc.gov/nchs/ahcd/ahcd_data_collection.htm#nhamcs_collection. Accessed Jul 23, 2012.
- 17A reason for visit classification for ambulatory care. Vital Health Stat 2. 1979;i-63., , .
- 18Asymptomatic bacteriuria--clinical significance and management. Nephrol Dial Transplant. 2001; 16(Suppl 6):135–6..
Data Supplement S1. Unadjusted analysis of ED visits due to urinary tract infection, 2001–2008.
Data Supplement S2. Results of multivariable logistic regression modeling for prediction of fever, altered mental status, and urinary tract symptoms among ED patients diagnosed as having urinary tract infection, 2001–2008.
Please note: Wiley Periodicals Inc. is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
|ACEM_1452_sm_DataSupplementS1.pdf||51K||Supporting info item|
|ACEM_1452_sm_DataSupplementS2.pdf||51K||Supporting info item|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.