The Effects of Patterned Urge-Response Toileting (PURT) on Urinary Incontinence among Nursing Home Residents Joyce Colling, PhD, RN, F.A.A.N.,* Joseph Ouslander, MD,t Betty Jo Hadley, RN, PhD,S Joan Eisch, RN, MS, FNP,§ and Emily Campbell, RN, MSII Objective: To test an individualized form of habit training for urinary incontinence (UI) among long-stay cognitively and/or physically impaired elderly nursing home residents over time. Design: Randomization of subjectsoccurred by nursing home unit. Baseline wet checks were done hourly for one 24-hour period at 3-week intervals over 12 weeks followed by 72 hours of continuous electronicmonitoring to establish precise voiding patterns for each subject. The 12-week intervention period was administered by indigenous staff after they attended a 4-hour UI educational program. Subjects were followed an additional 12 weeks to determine the extent of maintenance of the intervention among staff and subjects. Setting: Four non-profit nursing homes; west, mid-west, east coast. Patients: Consent was obtained from 154 (71%)who met primary inclusion criteria. Forty-one failed the secondary inclusion criteria leaving 113 who entered the 37-week study. Eighty-eight completed the study (experimental= 51, control
= 37);all were physically and/or mentally impaired, averaged age 85, and had either urge or urge/stress UI. Results: UI was significantly decreased during the 3-month period (P < 0.001). Eighty-six per cent showed improvement over baseline while one-third improved 25% or more over their baseline UI rate. The control group's UI increased during the same period of time. The volume of UI among the experimental group also decreased (P < 0.005) while the control group's UI volume increased. Conclusions: The training program was effective in reducing UI though compliance among nursing staff averaged only 70% of the prescribed toileting times. The success of this approach is similar to other recently described behavioral programs but achieved the reduction using only regular nursing staff. This individualized approach supports the recent regulatory thrust to individualize care to promote and maintain functional abilities and autonomy. J Am Geriatr SOC 39135-141,1992
he prevalence and costs of urinary incontinence (UI) among nursing home residents are well documented.'-' In addition, nursing staff report that UI is a difficult, frustrating, and onerous condition to manage.'-'' A variety of behavioral techniques have been developed to manage UI, including habit training, bladder retraining, vaginal cones, and pelvic muscle exercises, often in combination with biofeedback. These techniques for treating urinary incontinence involve the analysis and alteration of the relationship between the persons' symptoms and the maladaptive behavior. Most behavioral treatments involve three primary components: an educational program, scheduled voiding, and positive reinforcement for successful toileting. While,scattered reports of small-scale behavioral intervention studies for treating UI in nursing homes (NH) have appeared in the literature since the mid 1960's, the first large-scale clinical trials were com-
pleted within the last 3 years."-14 Each of these trials used a fixed 1-to-3-hour time interval for offering toileting assistance in conjunction with a variety of other more subtle behavioral techniques. In addition, specially trained research staff were used to carry out these interventions during subjects' usual waking hours only. The clinical trial for UI treatment in this report differs from previously studied behavioral interventions in several ways. First, an electronic data logger was used to obtain the exact times of voiding. Voidings occurring within 1-hour time blocks were grouped to develop individualized toileting schedules rather than subject all residents to the same fixed schedule. Second, the intervention, a prescribed toileting program called patterned urge-response toileting (PURT), was carried out 24 hours per day rather than only during waking hours. Third, indigenous nursing staff at the four nursing homes in the study, not research staff, carried out the toileting program after receiving a 4-hour specially designed educational program. Thus, PURT is a refinement of habit training since it emphasizes an individual patterned voiding schedule. While the primary purpose of this study was to determine the effect of PURT on frequency and volume of UI among nursing home residents, other objectives were to determine nursing staffs' ability and willingness to carry out a specific toileting program as well as to examine the effects of PURT on various physical, behavioral, and cost parameters. The emphasis in this
T
From the 'Community Health Care Systems Department, School of Nursing, Oregon Health Sciences University, Portland Oregon; tVictory Village, Jewish Homes for the Aging of Greater Los Angeles; MulticampusDivision of Geriatric Medicine and Gerontology, UCLA School of Medicine, Los Angeles, California; $University of Cincinnati, Cincinnati, Ohio; §State University of New YorkBinghamton, New York; JIUniversityof Wisconsin-Madison, Wisconsin. Supported by NIH-NCNR (R01-NR01554). Based on a paper presented at the Gerontology Society of America Annual Meeting, 1989. Address Correspondence to Dr. Joyce Colling, PhD, RN, FAAN, Professor of Nursing, Community Health Care Systems Department, School of Nursing, Oregon Health Sciences University (CB-SN), 3181 Southwest Sam Jackson Park Road, Portland, OR 97201
JAGS 40:135-141, 1992 0 1992 by the American Geriatrics Society
0002-8614/92/$3.50
136
COLLING ET AL
report, however, is on the primary objective of the study. Other aspects of the study will be published e1~ewhere.I~
METHODS Overview The study used a quasi-experimental repeated measures design consisting of a 12-weekbaseline period, a short period of time to obtain patterning data from the electronic data loggers, a 12-week intervention, and a 12-week post-intervention period. Frequency and volume of UI were measured by research staff for fourteen 24-hour periods of time, at 3-week intervals. The extended baseline period was necessary since there were no data to rule out the possible effect that increased attention alone in checking for UI may have an intervention effect. Further, there were little data to suggest the optimum length for intervention effect if one was to occur, nor if the intervention could be sustained after the treatment period. Within each of the four NHs, participating nursing units were designated as either experimental or control. Each unit consisted of 22 to 28 long-stay residents under the management of a registered nurse. Control and experimental units were assigned on the basis of geographical distance from each other in the facilities. Nursing staff were not exchanged between experimental and control units during the study by agreement with the facility administrators. It was not possible to randomize individual subjects to the experimental and control groups because the intervention was implemented by indigenous staff, and having both experimental and control subjects on the same unit would have led to confusion and blurring of the intervention with usual care. In addition, moving subjects was also not possible and could have created a confounding effect from a change in residents’ usual physical environment. Facilities Of the four participating NHs, which varied in size from 120 to 342 residents, three were in urban areas. All were non-profit facilities, but with differing management structures, and all cared for residents with moderate to severe functional impairments. Subjects Of the 763 residents in the four NHs, 415 met the initial eligibility criteria of age 65 or older and 3 or more episodes of involuntary loss of urine per week for the last 2 weeks as identified by nursing staff. Of this number, 218 met the additional criteria as capable of either giving their name or place of residence or toileting with the assistance of no more than one nursing staff member. Consent was obtained from 154 of the 218 eligible residents. A power analysis determined that 50 subjects per group would be sufficient to detect a difference in group means that accounted for at least 10% of the variance in the outcome variable. The power estimates were .85 and .73 for the respective .05 and .01 significance levels, which were conservative estimates because they did not take into account the statistical control gained by a repeated measures design.
IAGS-FEBRUARY 1992-VOL. 40, NO. 2
Assessment Procedures Assessment procedures included: (1) a detailed medical record review; (2) an assessment of mental, functional, and behavioral status; (3) a focused physical examination; (4)a urinalysis and culture and sensitivity; and (5) a simple urodynamic evaluation. Mental function was assessed using the 10-point SPMSQ.l6Activities of daily living were assessed using the Katz ADL Index.I7 The physical examination included a screening neurological examination as well as pelvic and rectal examinations. All subjects who had growth of lo5 or greater of a urinary pathogen from a urine culture were treated with an appropriate antimicrobial, and bacteriuria were eradicated before the simplified urodynamic evaluation and baseline data collection. The simplified urodynamic evaluation included a post-void residual determination, a stress test (forceful coughing) to detect stress incontinence, and simple cystometry to determine bladder capacity and the presence or absence of involuntary bladder contractions. Procedures for these tests have been described in detail elsewhere.”, Another part of the assessment involved a UI history and the determination of individual voiding patterns. Establishing patterns of incontinence has been discussed in previous studies using a voiding diary completed by the incontinent individual or by nursing staff who check for wetness every hour if the individual is unable to self-report. These methods are imprecise. In this study, an electronic data logger was added to obtain precise data on voiding patterns. This method was validated by analyzing 393 voiding events from randomly selected subjects over a total of 62 days of monitoring. Data logger and Incontinence Monitoring Record (IMR)” data from research assistants’ hourly wet checks resulted in correct identification of 98% of the incontinent events, with the further benefit of knowing precisely when voiding occurred. Both the IMR and data logger were used to obtain voiding data for 72 hours for each subject. The electronically generated data were then plotted on graph paper showing the exact time each void occurred. The mean time of voids encompassed within an hour’s block of time during the 72 hours was then computed. Regularity of voiding over the 3 days was established for an average of 85% of the voiding events for all subjects within hourly blocks of time. There was, however, variation among subjects as expected. A sample of these data is shown in Figure 1. Consistency of individual voiding patterns was tested in a sample of eight subjects by obtaining additional voiding data 3 months after obtaining the first set of data. Voiding times for each individual were remarkably similar over time with 82% of the voids occurring within the same hourly blocks of times. This regularity may be influenced by stable environmental factors such as regular times for meals, fluid intake, medications, activities, rest and sleep.”, 22 Individualized toileting prescriptions were constructed from these data and provided to facility staff to use on a daily basis for the 12-week intervention period. Specifically, staff were instructed to toilet individual subjects within 30 minutes prior to the mean
IAGS-FEBRUARY 1992-VOL. 40, NO. 2
PURT PROGRAM FOR URINARY INCONTINENCE
137
ate the effects of PURT on the urinary incontinence frequency ratio (UIFR) and the urinary incontinence volume ratio (UIVR).
RESULTS ible3n time or
voidang times ~ndicxedby dots Inside column
FIGURE 1. Voiding events recorded during 72-hour monitoring. Clinical nurse specialists who were part of the research staff plotted the data obtained from the Rustrak Ranger recorded over 72-hour periods for each experimental subject in order to determine patterns and an optimal toileting prescription. An example for one subject is shown above.
time of voids within an hour’s block of time. For instance, if the mean time of three voiding events during the 3 days of monitoring within a particular hour’s block of time was 10:15, then staff were instructed to toilet the resident between 9:45 and 10:15. All nursing staff on the experimental units participated in a 4-hour inservice UI education program which was taught by project staff. The program included causes and consequences of UI and detailed instructions on how to implement the PURT program. Staff were not paid for their participation in the program, but facilities were reimbursed for staff time taken by the education program. Pre and post tests were given to determine staffs’ level of knowledge of U1 and the PURT p r ~ g r a m . ’ ~ During the 12-week treatment period, project staff provided indigenous NH staff with voiding records on which subjects’ individualized prescribed toileting times were highlighted and concise instructions for carrying out PURT. Project staff then collected the daily voiding records during the 12-week treatment period and constructed weekly graphs of toileting results for each subject. These graphs were provided to facility staff as a visual record of their collective success for each subject for the current week as well as showing a comparison of the rate of success for the previous weeks. Further, throughout the intervention, project staff provided verbal encouragement to facility nursing staff to comply with the PURT prescriptions. Staff compliance was determined throughout the 12-week treatment period by research staff who observed staff on all shifts and days of the week and intermittently recorded staff compliance of the PURT toileting prescriptions on all subjects on compliance data forms. Outcome Measures The frequency and volume of all continent and incontinent voids were measured by project staff for 14 24-hour data collection periods spaced at 3-week intervals. During each 24-hour data collection period, project staff checked subjects for wetness/dryness each hour around the clock but did not initiate toileting. Urine volumes were recorded by measuring voided volumes using a commode hat for continent voids and by re-weighing pre-weighed disposable garments for incontinent voids using an Ohaus Electronic Scale. A 2 x 14, groups-by-time, repeated measures analysis of variance was employed to evalu-
Of the 154 residents in the four nursing homes for whom consent was obtained, 41 were excluded during the screening and assessment procedures because they failed to meet secondary exclusion criteria (see Appendix A for exclusion criteria). One hundred-thirteen residents were entered into the study, 63 in the experimental group and 50 in the control group. Twentyfive (22%) [12 (19%)of the experimental group and 13 (26%)of the control group] of the 113 subjects did not complete the 9-month study. Almost all of the 25 subjects who did not complete the study left because of acute illness or death. Subjects assigned to the experimental group who did not complete the study tended to be older (P =
= 37);all were physically and/or mentally impaired, averaged age 85, and had either urge or urge/stress UI. Results: UI was significantly decreased during the 3-month period (P < 0.001). Eighty-six per cent showed improvement over baseline while one-third improved 25% or more over their baseline UI rate. The control group's UI increased during the same period of time. The volume of UI among the experimental group also decreased (P < 0.005) while the control group's UI volume increased. Conclusions: The training program was effective in reducing UI though compliance among nursing staff averaged only 70% of the prescribed toileting times. The success of this approach is similar to other recently described behavioral programs but achieved the reduction using only regular nursing staff. This individualized approach supports the recent regulatory thrust to individualize care to promote and maintain functional abilities and autonomy. J Am Geriatr SOC 39135-141,1992
he prevalence and costs of urinary incontinence (UI) among nursing home residents are well documented.'-' In addition, nursing staff report that UI is a difficult, frustrating, and onerous condition to manage.'-'' A variety of behavioral techniques have been developed to manage UI, including habit training, bladder retraining, vaginal cones, and pelvic muscle exercises, often in combination with biofeedback. These techniques for treating urinary incontinence involve the analysis and alteration of the relationship between the persons' symptoms and the maladaptive behavior. Most behavioral treatments involve three primary components: an educational program, scheduled voiding, and positive reinforcement for successful toileting. While,scattered reports of small-scale behavioral intervention studies for treating UI in nursing homes (NH) have appeared in the literature since the mid 1960's, the first large-scale clinical trials were com-
pleted within the last 3 years."-14 Each of these trials used a fixed 1-to-3-hour time interval for offering toileting assistance in conjunction with a variety of other more subtle behavioral techniques. In addition, specially trained research staff were used to carry out these interventions during subjects' usual waking hours only. The clinical trial for UI treatment in this report differs from previously studied behavioral interventions in several ways. First, an electronic data logger was used to obtain the exact times of voiding. Voidings occurring within 1-hour time blocks were grouped to develop individualized toileting schedules rather than subject all residents to the same fixed schedule. Second, the intervention, a prescribed toileting program called patterned urge-response toileting (PURT), was carried out 24 hours per day rather than only during waking hours. Third, indigenous nursing staff at the four nursing homes in the study, not research staff, carried out the toileting program after receiving a 4-hour specially designed educational program. Thus, PURT is a refinement of habit training since it emphasizes an individual patterned voiding schedule. While the primary purpose of this study was to determine the effect of PURT on frequency and volume of UI among nursing home residents, other objectives were to determine nursing staffs' ability and willingness to carry out a specific toileting program as well as to examine the effects of PURT on various physical, behavioral, and cost parameters. The emphasis in this
T
From the 'Community Health Care Systems Department, School of Nursing, Oregon Health Sciences University, Portland Oregon; tVictory Village, Jewish Homes for the Aging of Greater Los Angeles; MulticampusDivision of Geriatric Medicine and Gerontology, UCLA School of Medicine, Los Angeles, California; $University of Cincinnati, Cincinnati, Ohio; §State University of New YorkBinghamton, New York; JIUniversityof Wisconsin-Madison, Wisconsin. Supported by NIH-NCNR (R01-NR01554). Based on a paper presented at the Gerontology Society of America Annual Meeting, 1989. Address Correspondence to Dr. Joyce Colling, PhD, RN, FAAN, Professor of Nursing, Community Health Care Systems Department, School of Nursing, Oregon Health Sciences University (CB-SN), 3181 Southwest Sam Jackson Park Road, Portland, OR 97201
JAGS 40:135-141, 1992 0 1992 by the American Geriatrics Society
0002-8614/92/$3.50
136
COLLING ET AL
report, however, is on the primary objective of the study. Other aspects of the study will be published e1~ewhere.I~
METHODS Overview The study used a quasi-experimental repeated measures design consisting of a 12-weekbaseline period, a short period of time to obtain patterning data from the electronic data loggers, a 12-week intervention, and a 12-week post-intervention period. Frequency and volume of UI were measured by research staff for fourteen 24-hour periods of time, at 3-week intervals. The extended baseline period was necessary since there were no data to rule out the possible effect that increased attention alone in checking for UI may have an intervention effect. Further, there were little data to suggest the optimum length for intervention effect if one was to occur, nor if the intervention could be sustained after the treatment period. Within each of the four NHs, participating nursing units were designated as either experimental or control. Each unit consisted of 22 to 28 long-stay residents under the management of a registered nurse. Control and experimental units were assigned on the basis of geographical distance from each other in the facilities. Nursing staff were not exchanged between experimental and control units during the study by agreement with the facility administrators. It was not possible to randomize individual subjects to the experimental and control groups because the intervention was implemented by indigenous staff, and having both experimental and control subjects on the same unit would have led to confusion and blurring of the intervention with usual care. In addition, moving subjects was also not possible and could have created a confounding effect from a change in residents’ usual physical environment. Facilities Of the four participating NHs, which varied in size from 120 to 342 residents, three were in urban areas. All were non-profit facilities, but with differing management structures, and all cared for residents with moderate to severe functional impairments. Subjects Of the 763 residents in the four NHs, 415 met the initial eligibility criteria of age 65 or older and 3 or more episodes of involuntary loss of urine per week for the last 2 weeks as identified by nursing staff. Of this number, 218 met the additional criteria as capable of either giving their name or place of residence or toileting with the assistance of no more than one nursing staff member. Consent was obtained from 154 of the 218 eligible residents. A power analysis determined that 50 subjects per group would be sufficient to detect a difference in group means that accounted for at least 10% of the variance in the outcome variable. The power estimates were .85 and .73 for the respective .05 and .01 significance levels, which were conservative estimates because they did not take into account the statistical control gained by a repeated measures design.
IAGS-FEBRUARY 1992-VOL. 40, NO. 2
Assessment Procedures Assessment procedures included: (1) a detailed medical record review; (2) an assessment of mental, functional, and behavioral status; (3) a focused physical examination; (4)a urinalysis and culture and sensitivity; and (5) a simple urodynamic evaluation. Mental function was assessed using the 10-point SPMSQ.l6Activities of daily living were assessed using the Katz ADL Index.I7 The physical examination included a screening neurological examination as well as pelvic and rectal examinations. All subjects who had growth of lo5 or greater of a urinary pathogen from a urine culture were treated with an appropriate antimicrobial, and bacteriuria were eradicated before the simplified urodynamic evaluation and baseline data collection. The simplified urodynamic evaluation included a post-void residual determination, a stress test (forceful coughing) to detect stress incontinence, and simple cystometry to determine bladder capacity and the presence or absence of involuntary bladder contractions. Procedures for these tests have been described in detail elsewhere.”, Another part of the assessment involved a UI history and the determination of individual voiding patterns. Establishing patterns of incontinence has been discussed in previous studies using a voiding diary completed by the incontinent individual or by nursing staff who check for wetness every hour if the individual is unable to self-report. These methods are imprecise. In this study, an electronic data logger was added to obtain precise data on voiding patterns. This method was validated by analyzing 393 voiding events from randomly selected subjects over a total of 62 days of monitoring. Data logger and Incontinence Monitoring Record (IMR)” data from research assistants’ hourly wet checks resulted in correct identification of 98% of the incontinent events, with the further benefit of knowing precisely when voiding occurred. Both the IMR and data logger were used to obtain voiding data for 72 hours for each subject. The electronically generated data were then plotted on graph paper showing the exact time each void occurred. The mean time of voids encompassed within an hour’s block of time during the 72 hours was then computed. Regularity of voiding over the 3 days was established for an average of 85% of the voiding events for all subjects within hourly blocks of time. There was, however, variation among subjects as expected. A sample of these data is shown in Figure 1. Consistency of individual voiding patterns was tested in a sample of eight subjects by obtaining additional voiding data 3 months after obtaining the first set of data. Voiding times for each individual were remarkably similar over time with 82% of the voids occurring within the same hourly blocks of times. This regularity may be influenced by stable environmental factors such as regular times for meals, fluid intake, medications, activities, rest and sleep.”, 22 Individualized toileting prescriptions were constructed from these data and provided to facility staff to use on a daily basis for the 12-week intervention period. Specifically, staff were instructed to toilet individual subjects within 30 minutes prior to the mean
IAGS-FEBRUARY 1992-VOL. 40, NO. 2
PURT PROGRAM FOR URINARY INCONTINENCE
137
ate the effects of PURT on the urinary incontinence frequency ratio (UIFR) and the urinary incontinence volume ratio (UIVR).
RESULTS ible3n time or
voidang times ~ndicxedby dots Inside column
FIGURE 1. Voiding events recorded during 72-hour monitoring. Clinical nurse specialists who were part of the research staff plotted the data obtained from the Rustrak Ranger recorded over 72-hour periods for each experimental subject in order to determine patterns and an optimal toileting prescription. An example for one subject is shown above.
time of voids within an hour’s block of time. For instance, if the mean time of three voiding events during the 3 days of monitoring within a particular hour’s block of time was 10:15, then staff were instructed to toilet the resident between 9:45 and 10:15. All nursing staff on the experimental units participated in a 4-hour inservice UI education program which was taught by project staff. The program included causes and consequences of UI and detailed instructions on how to implement the PURT program. Staff were not paid for their participation in the program, but facilities were reimbursed for staff time taken by the education program. Pre and post tests were given to determine staffs’ level of knowledge of U1 and the PURT p r ~ g r a m . ’ ~ During the 12-week treatment period, project staff provided indigenous NH staff with voiding records on which subjects’ individualized prescribed toileting times were highlighted and concise instructions for carrying out PURT. Project staff then collected the daily voiding records during the 12-week treatment period and constructed weekly graphs of toileting results for each subject. These graphs were provided to facility staff as a visual record of their collective success for each subject for the current week as well as showing a comparison of the rate of success for the previous weeks. Further, throughout the intervention, project staff provided verbal encouragement to facility nursing staff to comply with the PURT prescriptions. Staff compliance was determined throughout the 12-week treatment period by research staff who observed staff on all shifts and days of the week and intermittently recorded staff compliance of the PURT toileting prescriptions on all subjects on compliance data forms. Outcome Measures The frequency and volume of all continent and incontinent voids were measured by project staff for 14 24-hour data collection periods spaced at 3-week intervals. During each 24-hour data collection period, project staff checked subjects for wetness/dryness each hour around the clock but did not initiate toileting. Urine volumes were recorded by measuring voided volumes using a commode hat for continent voids and by re-weighing pre-weighed disposable garments for incontinent voids using an Ohaus Electronic Scale. A 2 x 14, groups-by-time, repeated measures analysis of variance was employed to evalu-
Of the 154 residents in the four nursing homes for whom consent was obtained, 41 were excluded during the screening and assessment procedures because they failed to meet secondary exclusion criteria (see Appendix A for exclusion criteria). One hundred-thirteen residents were entered into the study, 63 in the experimental group and 50 in the control group. Twentyfive (22%) [12 (19%)of the experimental group and 13 (26%)of the control group] of the 113 subjects did not complete the 9-month study. Almost all of the 25 subjects who did not complete the study left because of acute illness or death. Subjects assigned to the experimental group who did not complete the study tended to be older (P =
