|Home | About | Journals | Submit | Contact Us | Français|
David B. Reuben, MD, 10945 Le Conte Avenue, Suite 2339 Box 951687, Los Angeles, California 90095-1687, Phone: (310) 825-8253, dreuben/at/mednet.ucla.edu
Emmett Keeler, Ph.D., RAND Health Program, 1776 Main St, Santa Monica, CA 90407-2138, (310) 393-0411 x 7239, emmett/at/rand.org
David A. Ganz, MD Ph.D., VA Greater Los Angeles Healthcare System (11G), 11301 Wilshire Blvd., Building 220, Room 313, Los Angeles, CA 90073, Telephone: (310) 268-4110, Fax: (310) 268-4842, dganz/at/mednet.ucla.edu
Constance H. Fung, MD, MSHS, UCLA Multicampus Program in Geriatric Medicine and Gerontology, 10945 Le Conte Avenue, Suite 2339, Los Angeles, CA 90095-1687, Telephone (310) 825-8253, Fax (310) 794-2199, cfung/at/ucla.edu
Paul Shekelle MD, PhD, RAND Health, 1776 Main Street, Santa Monica, CA 90401, Telephone: (310) 393-0411, ext. 6669, Fax: 310) 393-4818, shekelle/at/rand.org
Carol P. Roth, RN, MPH, RAND Health, 1776 Main Street, Santa Monica, CA 90401, Telephone: (310) 393-0411, ext. 6425, Fax: 310) 393-4818, roth/at/rand.org
Neil S. Wenger, MD, UCLA Med-GIM & HSR, Box 951736, 911 Broxton Plaza, Los Angeles, CA 90095-1736, Phone 310-794-2288, Fax 310-794-0732, nwenger/at/mednet.ucla.edu
Care for falls and urinary incontinence (UI) among older patients is inadequate. One possible explanation is that physicians provide less recommended care to patients who are not as concerned about their falls and UI.
To test whether patient-reported severity for two geriatric conditions, falls and UI, is associated with quality of care.
Prospective cohort study of elders with falls and/or fear of falling (n=384) and UI (n=163).
Participants in the Assessing Care of Vulnerable Elders-2 Study (2002–3), which evaluated an intervention to improve the care for falls and UI among older (age ≥75) ambulatory care patients with falls/fear of falling or UI.
Falls Efficacy Scale (FES) and the Incontinence Quality of Life (IQOL) surveys measured at baseline, quality of care measured by a 13-month medical record abstraction.
There was a small difference in falls quality scores across the range of FES, with greater patient-perceived falls severity associated with better odds of passing falls quality indicators (OR 1.11 (95% CI 1.02–1.21) per 10-point increment in FES). Greater patient-perceived UI severity (IQOL score) was not associated with better quality of UI care.
Although older persons with greater patient-perceived falls severity receive modestly better quality of care, those with more distressing incontinence do not. For both conditions, however, even the most symptomatic patients received less than half of recommended care. Low patient-perceived severity of condition is not the basis of poor care for falls and UI.
Older patients often fail to receive recommended care for common geriatric conditions such as fear of falling, falls and urinary incontinence (UI) but we understand little about the factors associated with such deficits in care.1 Previous studies identified system2–5 and provider6–16 characteristics that contribute to variations in care quality.
Patient-related factors are another potential source of variation in care quality. Patients with more co-morbidities17–19 receive better quality of care, and patients needing time-consuming services such as counseling receive poorer quality of care.18 Unrelated co-morbidities and some sociodemographic characteristics may also negatively affect care.20–28
Some studies include illness severity as a predictor of better quality of care. Studies suggest that better care is delivered to those with objective clinical measures of severity: higher blood pressures are associated with better hypertension care;29, 30 lower CD4 counts are associated with better HIV care;31 and pneumonia severity measures are associated with pneumonia guideline adherence.32, 33 No studies, however, have focused on whether subjective patient report of disease severity is associated with better quality of care. We sought to determine whether subjective patient-reported severity of fear of falling, falls, and UI affected the quality of care provided.
This study used falls and UI data from the Assessing Care of Vulnerable Elders-2 (ACOVE-2) study,34 a controlled trial to improve the quality of care for falls, UI, and dementia among older outpatients. As part of this study, we aimed to analyze whether patient-reported disease severity was related to subsequent quality of care for falls, fear of falling, and UI. A self-reported dementia severity measure was unavailable for study.
The ACOVE-2 intervention implemented a practice re-design intervention35. Patients aged ≥75 years at intervention and control sites were screened by office personnel with yes/no screening questions concerning any fear of falling, serious falls (2+ falls or fall requiring physician attention in past year) and UI symptoms (bothersome enough to seek medical care).
Upon enrollment, those with fear of falling and/or falls were administered the 10-item Falls Efficacy Scale (FES),36 a measure associated with severity of future falls, gait and balance impairment, and disability.36–39 Patients who screened positively for fear of falling but denied falls were excluded if they had a “negative” FES (i.e., described no falling concern on all FES items). Those with bothersome UI answered the 22-item Incontinence Quality of Life (IQOL) survey, a measure of incontinence severity and quality of life.40, 41 Patients could be in both UI and falls samples.
We used ACOVE process-of-care quality indicators (QIs) to evaluate falls and UI care. The number of QIs triggered per participant varied, due to presence/absence of clinical criteria indicated by the QIs.
We applied these measures of quality to all outpatient primary care and specialist medical records for a 13-month period. If the patient received the recommended process, a score of 1 was awarded; if not, a score of 0 was assigned. Documentation of patient refusal received full credit. For selected QIs (Falls QIs 3–5 and UI QIs 1 and 6), individuals with advanced dementia or life expectancy ≤ 6 months were not evaluated.42 In addition, if the patient had already received a workup and/or completed recommended therapies for the falls or UI condition (i.e., “maximal treatment”), then QIs were excluded.
Our primary outcome was a dichotomous variable representing whether the triggered QI was passed or failed, clustered at the level of the patient.
The primary variable of interest for the falls analysis was the severity of patient’s concern for falls during daily activities (e.g., bathing, walking) measured by the FES.36 FES scores are calculated as the sum of responses (1 point for “not at all” concerned to 4 points for “very” concerned) for 10 activities, ranging from 10 to 40. To facilitate interpretation of the results, we rescaled the FES score from 0 to 100 points (higher=more severe concern). Those with a re-scaled score ≥25 have severe fear of falling, a threshold associated with increased risk of future falls.39 Our analytic sample only included subjects with ≤2 missing FES items.
The primary variable of interest for the UI analysis was severity of distress and bother (e.g., limitations in daily activities, embarrassment, or psychosocial consequences) caused by UI symptoms as measured by the 22-item IQOL survey.41 Responses to each item are graded on a 5-point scale ranging from “extremely” (1 point) to “not at all” (5 points). The IQOL is a sum of the responses, rescaled on a 100-point scale.41 For ease of comparison with falls analyses, we reversed the IQOL score so that higher scores reflected worse symptoms (i.e., “reversed-IQOL”, or “rIQOL” score). Those with rIQOL scores of ≤50 had mild-to-moderate UI,40 while scores ≥50 indicate moderate-to-severe disease.40 We excluded patients with >3 missing items.
For both analyses, we calculated full FES and rIQOL scores for those with few missing items (1–2 for FES, 1–3 for rIQOL) using multiple imputation.43 IQOL and FES scores obtained from proxy respondents were excluded because they overstate FES and understate IQOL compared to subjects by 0.9 and 0.7 standard deviations, respectively.44
We classified two falls QIs (# 4,5) and three UI QIs (# 4, 5, 6, Table 2) as “secondary” QIs because they could be triggered only by an abnormality (e.g., abnormal gait) identified as a result of passing another QI (e.g., gait examination). We considered this as a potential confounder because patients with more severe conditions might be more likely to trigger and pass a secondary QI because they passed the preceding primary QI.
We first performed unadjusted logistic regressions of the quality scores for falls QIs with FES scores and the quality scores for UI QIs with rIQOL scores. Then we used multivariable logistic regression to predict the odds of passing versus failing the QIs, controlling for age, gender, number of QIs triggered for all ACOVE conditions (proxy for co-morbidity), and primary versus secondary QI. Because the ACOVE-2 intervention improved the quality of care for falls and incontinence34, we controlled for the intervention and considered an interaction term with baseline FES or rIQOL. Standard errors were adjusted for clustering by patient, then by physician. Analyses were performed using STATA version 10.0.
Of 2671 screened older (age ≥75) patients, 53 (11%) reported falls, 369 (74%) reported fear of falling, and 78 (16%) reported both falls and fear of falling. After inclusion criteria were applied, our analytic sample for fear of falling/falls (“falls sample”) consisted of 384 (77%) patients evaluated for 656 QIs. There were 235 (9% of screened) patients who endorsed bothersome UI symptoms; 163 (69%) met inclusion criteria and were evaluated for 789 UI QIs (Figure 1).
Mean subject age was 81 years. The falls sample was 34% male; the UI sample was 20% male. The overall pass rate was 35% for the falls and 32% for the UI QIs (Table 1). Pass rates for individual QIs are displayed in Table 2.
In unadjusted models, higher (more symptomatic) FES scores were associated with higher odds of passing falls QIs (OR 1.09 for every 10-point increase in FES, p=.007).
In the multivariable logistic regression analyses, each 10-point increase in FES was associated with 1.11 (95% CI 1.02–1.21) times the odds of passing falls QIs (Table 3). Being in the intervention group increased odds of passing the falls QIs (OR=2.77, 95% CI 1.56–4.92) but did not modify the relationship between FES and quality of care (p=.6 for interaction term between FES and intervention group). Age, gender, co-morbidity, and primary versus secondary QI had no effect on quality.
The predicted probabilities of quality of falls care for a hypothetical patient across the range of FES scores were obtained from the full multivariable model. In the graphed example (Figure 2), the predicted probability of passing across the range of FES scores (0 to 100) increased from 18% (95% CI 13–24%) to 38% (95% CI 25–47%). Across the FES interquartile range, the predicted difference in passing was 7% (20% versus 27%). Figure 2 shows that even for those with the highest concern and probability of falling, recommended care was provided approximately one-third of the time.
In this study of subjective disease severity and quality of care, we found a small effect of concern about falling on quality of falls care (a 7% absolute difference in quality of care over the FES interquartile range). There was no relationship between distress due to UI symptoms and quality of care.
The mild effect of patient-reported fear of falling on falls care is consistent with prior research. Greater fear of falling is related to likelihood of falling,37–39 so physicians may have provided more recommended falls care to patients with higher FES scores to prevent falls in higher-risk patients with greater potential for benefit.36, 39 Higher patient-reported FES scores may reflect greater patient priority in receiving falls care, which is consistent with prior work suggesting that greater symptom severity is associated with healthcare-seeking behavior.45–47 Also, greater severity increases the likelihood of clinical detection.48
However, even patients with severe fear of falling or incontinence symptoms do not receive high quality of care. Thus, the degree of patient-reported severity for falls and UI is not the overall basis of poor care for these conditions.
Our findings have implications for the science of quality measurement. Process-of-care QIs, typically identify a single clinical threshold at which a particular care process should be performed for all eligible patients. In most cases, they represent a minimum standard of quality of care for a population. They do not typically require an increasing intensity of care in response to small increments in patient-perceived disease severity; hence, most process-of-care QIs do not adjust for severity. Finding little or no association between patient-perceived disease severity and performance provides support that adjustment for patient-perceived disease severity is unnecessary.
The minimal relationship between patient-reported disease severity and quality of care for falls and lack of relationship for UI is contrary to findings in hypertension, HIV, and pneumonia where objective clinical severity relates to measured quality of care. 31–33 This may be because clinicians in our study were not provided with FES and IQOL scores, whereas clinicians were likely aware of clinical disease severity in the cited studies. Additionally, we considered whether the screening questions identified patients with mild conditions; however, FES and IQOL scores suggest moderate-to-severe concern for falling and distressing UI, and other studies have used similar questions to screen for falls49 and fear of falling39 interventions. Last, UI and falls care may be considered less critical or more burdensome than care for conditions studied previously.
This study has several limitations. Although FES and IQOL scores are associated with objective clinical indicators of falls and UI severity (e.g., future falls36–39 and number of incontinent episodes per week40, 41), it is possible that they were inadequate proxies for overall clinical severity; better measures could be collected in the future. We did not measure other types of patient-reported variables that have been hypothesized to affect quality of care such as patients’ trust in their physician.50 Our sample lacked ethnic diversity. We lacked sample size to analyze each quality indicator separately. Finally, our results are not generalizable to older patients who require proxies due to cognitive impairment.
Despite these limitations, we were able to rigorously examine patient-reported disease severity using scales that differentiated our subjects on a continuum of severity, and the symptom surveys were administered prior to delivery of care in anticipation of this planned analysis. We believe that our results are generalizable to older community-dwelling elders because our study selected a sample based only on age and screening questions specific to the diseases studied.
In conclusion, this study found that patient-reported fall severity is weakly related to quality of care, and patient-reported UI severity is not associated with quality of care. Given the wide gap between falls care1, 34 and recommended care, the relationship between patient-perceived severity and quality is likely of little consequence.
Dr. Min is supported by the AHRQ (R21 HS017621). Dr. Min was on the UCLA faculty during the writing of this manuscript, supported by NIA-UCLA (K12 AG001004). Dr. Ganz is funded by the U.S. Department of Veterans Affairs, Veterans Health Administration, VA Health Services Research & Development (HSR&D) Service through the VA Greater Los Angeles HSR&D Center of Excellence (Project # VA CD2 08-012-1). This research was also supported by the UCLA Claude Pepper Older Americans Independence Center funded by the National Institute of Aging (AG028748).