Brent C. James, MD, MStat

Inpatient pediatrics is undergoing a paradigm shift in at least 3 ways. First, more children with chronic disease are being cared for in the hospital over time.1 Second, previous inpatient conditions are treated at home with advancing technology such as peripherally‐inserted catheters.2 Third, there are new areas of growing specialization, such as hospital medicine, in which the practitioners deliver more efficient care.3, 4

Nationwide, there is increasing pressure to improve inpatient quality of care. The Institute of Medicine defines 6 aims for improvement, including timeliness (reducing waits and sometimes harmful delays for both those who receive and those who give care) and efficiency of care (avoiding waste, including waste of equipment, supplies, ideas, and energy).5 Reducing unnecessary stays in the hospital is a potential quality measure that hospitals may use to address the timeliness and efficiency of care delivered to hospitalized children.

Delays in discharge have been used as markers of unnecessary stays in the hospital for inpatient adult and pediatric care,6, 7 but these are limited to inpatient systems from almost 20 years ago. Current reasons why patients are delayed from discharge, if at all, are not well described. We undertook this study to describe delays in hospital discharges at a tertiary‐care children's hospital in terms of number of patients, length of days of delay, and type of delay. In addition, we sought to characterize the impact of discharge delays on overall length of stay (LOS) and costs.

Methods

Results

During the 31 days of the study, 171 patients occupied hospital beds an average of 7.3 days on the 2 inpatient medical teams, for a total of 911 inpatient days. Seven patients were admitted prior to August 1; 6 of these were discharged during the month of August and 1 stayed through the entire month and was discharged in September. Three additional patients were admitted in August and discharged in September. There were 6 readmissions during the month of August, and 1 patient was excluded from the study because of lack of sufficient information. All patients with delays were able to be classified according to the Delay Tool taxonomy. Interrater reliability for the 2 study physicians was 98%.

The characteristics between the patients who did and did not experience a delay in discharge are shown in Table 1. Thirty‐nine of 171 patients (22.8%), experienced at least 1 delay day. Eighteen of 39 patients had only 1 delay day (46.2%) and 11 patients experienced 2 delayed days (28.2%) (Figure 1). The average length of delay was 2.1 days.

Figure 1

Delays attributed to physician responsibility accounted for 42.3% (16.5/39) of patient delays (conservative management or clinical decision‐making), with discharge planning delays accounting for 21.8% (family‐related, patient‐related, and hospital‐related problems), consultation for 14.1% (delay in obtaining or lack of follow‐up), test scheduling for 12.8%, and obtaining test results for 5.1% (ordering and weekend scheduling). There were no primary delays due to surgery, education and research, or unavailability of outside resources such as a skilled nursing bed. Four patients had a single additional secondary cause of delay assigned to them, related to physician responsibility, consultation, surgery and test scheduling; these were split, attributing 0.5 patients to each delay type (thus, the 17/39 patients delayed for physician responsibility was analyzed as 16.5/39) (Table 2).

There were 82 delay‐related hospital days of 911 total inpatient days on the 2 medical teams for August 2004 (9%). More than $170,000 in excess costs was incurred due to delay days from a total of approximately 1.9 million dollars in patient costs for the month (8.9%).

Discussion

This study finds that discharge delays in a tertiary care children's hospital are common; almost 1 in 4 patients experienced a medically unnecessary excess hospital stay of at least 1 day. The average length of a delay was 2.1 days, and overall, delays consumed 9% of pediatric hospital days and 8.9% of total costs. The most common reason for a delay was related to physician clinical care, including excessively conservative management and variability in clinical decision‐making.

Our study results are similar to the other 2 published studies that use the Delay Tool. In the adult and pediatric studies, between 10% and 30% of patients experienced a delay in discharge, with the average length of delay between 2.9 and 3 days.6, 7 Although both studies were conducted at teaching hospitals, what is particularly interesting is that they were conducted almost 20 years ago. During this period of time, there has been a shift in the inpatient pediatric patient population. In recent years, children who are cared for in the hospital have more chronic illnesses.1 In addition, there has been a shift in the types of conditions that may be cared for at home and those that now require inpatient stay.2 Despite this, delays continue at a similar proportion, but the cause of delays have shifted from scheduling and consultation to physician responsibility.

There is another tool in the literature which is more widely used, the Pediatric Appropriateness Evaluation Protocol (PAEP), which is based on the Appropriateness Evaluation Protocol for adults.1417 This tool is used to determine the appropriateness of ongoing hospitalization, not the cause of delay if ongoing hospitalization is inappropriate. The 3 areas that are evaluated (medical services, nursing and ancillary services, and patient's condition) have objective criteria that dictate if the hospitalization is appropriate or not (eg, parenteral (intravenous) therapy for at least 8 hours on that day, under nursing and ancillary services). The PAEP may be less sensitive given today's healthcare resource utilization climate. Many clinicians and families would agree that insertion of a peripheral central catheter is an acceptable form of outpatient treatment for many pediatric conditions. In conjunction with the Delay Tool, the PAEP could be used to determine if a delay occurred, then the Delay Tool used to categorize the cause of the delay. We choose to use expert clinician judgment to determine if a delay had occurred. We were more interested in why patients who are admitted (appropriately or inappropriately) cannot be discharged sooner, thus allowing for future intervention studies targeted to impact delays in discharge, as elucidated in this study. The Delay Tool specifically allowed us to categorize the reasons for delays. Given that the average LOS for patients in the nondelayed group was over 4 days, despite not using a tool such as the PAEP, we believe that these were likely to be appropriate admissions.

A recent study reported the first use of the Medical Care Appropriateness Protocol (MCAP) in a tertiary‐care children's hospital. The authors used the MCAP to determine the impact of an intervention on reducing inappropriate hospitals days for children. This tool is similarly labor‐intensive to the Delay Tool. Interestingly, this Canadian study found a high rate of inappropriate hospital days (47%), which may be in part attributable to a different outcome measurement tool and/or a different health care system.18

There are several limitations to our study that deserve mention. The Delay Tool requires clinician judgment regarding whether or not there was a delay in discharge for that day. We may have introduced some bias in our study, as hospitalist investigators assigned the delay and blinding to the attending physician specialty of record was not feasible. However, our results are similar to the other 2 published studies that have used this tool, and we specifically chose not to analyze or report results in terms of hospitalist and nonhospitalist attending physicians. The Delay Tool is not designed to differentiate shorter delays in terms of hours instead of days (eg, due to the inability for the patient to get a ride home). Shorter delays may be of particular importance depending on the occupancy rate of the hospital, the demand for beds, and other patient and hospital factors. We could not capture these shorter delays (although they did occur frequently) due to the original description of the Delay Tool. In addition, we would not have been able to report data on the impact on LOS and costs, as these are attributed to whole days in the hospital. However, if we had been able to differentiate shorter delays, this would bias our results to show a greater percentage of delays over smaller increments of time. Generalizability is an issue, given that this was a single‐center study. This study sample included over 80 different attending physicians participating in community pediatrician, subspecialty, and hospitalist practice groups. However, the patient population at PCMC is similar to other medium and large children's hospitals in the United States. The month observed may not reflect the entire year of hospitalizationsthere may be seasonal variations with delays depending on the volume and type of illness seen. The study was conducted in August, when there are newer house staff present. However, physician responsibility, which was the largest source of delays in our study, had little attribution to house staff. Most of the decisions were those of attending physicians, which would largely be unaffected by the time of year of the study. Finally, we were unable to assess the safety of the potential earlier discharge, as this was an observational study. However, in any future intervention studies examining processes to discharge patients sooner, measures of safety to the patient are a necessity. Finally, given the potential of ongoing admission, even on the date of discharge of our most fragile patients, this approach to discovering causes of delay may not apply to this important group, which is responsible for significant and growing resource utilization.

Despite these limitations, our findings demonstrate that in an era of children staying in the hospital less, and more medically‐complex children being admitted,10 a substantial number of children who are hospitalized at a children's hospital may have been discharged sooner. The majority of these decisions were directly related to physician responsibility. As consumers, providers, and hospitals work together to develop quality measures that are reflective of inpatient pediatric care, the Delay Tool may be able to highlight 2 aims of quality (ie, timeliness and efficiency of care) that could be used to assess the impact of interventions designed to safely discharge patients sooner. Interventions such as audit‐feedback,18 clinical guideline deployment,19 and hospitalist systems of care4 continue to hold the promise of earlier discharge; however, tools designed to measure inappropriate use of hospital days should be employed to demonstrate their effectiveness. Our study demonstrates ongoing waste in children's hospitals.

Conclusions

Almost 1 out of 4 patients in this 1‐month period could have been discharged sooner than they were. The impact of delays on costs and LOS are substantial and should provide strong incentives to develop effective interventions. Such interventions will need to address variations in physician criteria for discharge, more efficient discharge planning, and timely scheduling of consultation and diagnostic testing.

Inpatient pediatrics is undergoing a paradigm shift in at least 3 ways. First, more children with chronic disease are being cared for in the hospital over time.1 Second, previous inpatient conditions are treated at home with advancing technology such as peripherally‐inserted catheters.2 Third, there are new areas of growing specialization, such as hospital medicine, in which the practitioners deliver more efficient care.3, 4

Nationwide, there is increasing pressure to improve inpatient quality of care. The Institute of Medicine defines 6 aims for improvement, including timeliness (reducing waits and sometimes harmful delays for both those who receive and those who give care) and efficiency of care (avoiding waste, including waste of equipment, supplies, ideas, and energy).5 Reducing unnecessary stays in the hospital is a potential quality measure that hospitals may use to address the timeliness and efficiency of care delivered to hospitalized children.

Delays in discharge have been used as markers of unnecessary stays in the hospital for inpatient adult and pediatric care,6, 7 but these are limited to inpatient systems from almost 20 years ago. Current reasons why patients are delayed from discharge, if at all, are not well described. We undertook this study to describe delays in hospital discharges at a tertiary‐care children's hospital in terms of number of patients, length of days of delay, and type of delay. In addition, we sought to characterize the impact of discharge delays on overall length of stay (LOS) and costs.

Methods

Results

During the 31 days of the study, 171 patients occupied hospital beds an average of 7.3 days on the 2 inpatient medical teams, for a total of 911 inpatient days. Seven patients were admitted prior to August 1; 6 of these were discharged during the month of August and 1 stayed through the entire month and was discharged in September. Three additional patients were admitted in August and discharged in September. There were 6 readmissions during the month of August, and 1 patient was excluded from the study because of lack of sufficient information. All patients with delays were able to be classified according to the Delay Tool taxonomy. Interrater reliability for the 2 study physicians was 98%.

The characteristics between the patients who did and did not experience a delay in discharge are shown in Table 1. Thirty‐nine of 171 patients (22.8%), experienced at least 1 delay day. Eighteen of 39 patients had only 1 delay day (46.2%) and 11 patients experienced 2 delayed days (28.2%) (Figure 1). The average length of delay was 2.1 days.

Figure 1

Delays attributed to physician responsibility accounted for 42.3% (16.5/39) of patient delays (conservative management or clinical decision‐making), with discharge planning delays accounting for 21.8% (family‐related, patient‐related, and hospital‐related problems), consultation for 14.1% (delay in obtaining or lack of follow‐up), test scheduling for 12.8%, and obtaining test results for 5.1% (ordering and weekend scheduling). There were no primary delays due to surgery, education and research, or unavailability of outside resources such as a skilled nursing bed. Four patients had a single additional secondary cause of delay assigned to them, related to physician responsibility, consultation, surgery and test scheduling; these were split, attributing 0.5 patients to each delay type (thus, the 17/39 patients delayed for physician responsibility was analyzed as 16.5/39) (Table 2).

There were 82 delay‐related hospital days of 911 total inpatient days on the 2 medical teams for August 2004 (9%). More than $170,000 in excess costs was incurred due to delay days from a total of approximately 1.9 million dollars in patient costs for the month (8.9%).

Discussion

This study finds that discharge delays in a tertiary care children's hospital are common; almost 1 in 4 patients experienced a medically unnecessary excess hospital stay of at least 1 day. The average length of a delay was 2.1 days, and overall, delays consumed 9% of pediatric hospital days and 8.9% of total costs. The most common reason for a delay was related to physician clinical care, including excessively conservative management and variability in clinical decision‐making.

Our study results are similar to the other 2 published studies that use the Delay Tool. In the adult and pediatric studies, between 10% and 30% of patients experienced a delay in discharge, with the average length of delay between 2.9 and 3 days.6, 7 Although both studies were conducted at teaching hospitals, what is particularly interesting is that they were conducted almost 20 years ago. During this period of time, there has been a shift in the inpatient pediatric patient population. In recent years, children who are cared for in the hospital have more chronic illnesses.1 In addition, there has been a shift in the types of conditions that may be cared for at home and those that now require inpatient stay.2 Despite this, delays continue at a similar proportion, but the cause of delays have shifted from scheduling and consultation to physician responsibility.

There is another tool in the literature which is more widely used, the Pediatric Appropriateness Evaluation Protocol (PAEP), which is based on the Appropriateness Evaluation Protocol for adults.1417 This tool is used to determine the appropriateness of ongoing hospitalization, not the cause of delay if ongoing hospitalization is inappropriate. The 3 areas that are evaluated (medical services, nursing and ancillary services, and patient's condition) have objective criteria that dictate if the hospitalization is appropriate or not (eg, parenteral (intravenous) therapy for at least 8 hours on that day, under nursing and ancillary services). The PAEP may be less sensitive given today's healthcare resource utilization climate. Many clinicians and families would agree that insertion of a peripheral central catheter is an acceptable form of outpatient treatment for many pediatric conditions. In conjunction with the Delay Tool, the PAEP could be used to determine if a delay occurred, then the Delay Tool used to categorize the cause of the delay. We choose to use expert clinician judgment to determine if a delay had occurred. We were more interested in why patients who are admitted (appropriately or inappropriately) cannot be discharged sooner, thus allowing for future intervention studies targeted to impact delays in discharge, as elucidated in this study. The Delay Tool specifically allowed us to categorize the reasons for delays. Given that the average LOS for patients in the nondelayed group was over 4 days, despite not using a tool such as the PAEP, we believe that these were likely to be appropriate admissions.

A recent study reported the first use of the Medical Care Appropriateness Protocol (MCAP) in a tertiary‐care children's hospital. The authors used the MCAP to determine the impact of an intervention on reducing inappropriate hospitals days for children. This tool is similarly labor‐intensive to the Delay Tool. Interestingly, this Canadian study found a high rate of inappropriate hospital days (47%), which may be in part attributable to a different outcome measurement tool and/or a different health care system.18

There are several limitations to our study that deserve mention. The Delay Tool requires clinician judgment regarding whether or not there was a delay in discharge for that day. We may have introduced some bias in our study, as hospitalist investigators assigned the delay and blinding to the attending physician specialty of record was not feasible. However, our results are similar to the other 2 published studies that have used this tool, and we specifically chose not to analyze or report results in terms of hospitalist and nonhospitalist attending physicians. The Delay Tool is not designed to differentiate shorter delays in terms of hours instead of days (eg, due to the inability for the patient to get a ride home). Shorter delays may be of particular importance depending on the occupancy rate of the hospital, the demand for beds, and other patient and hospital factors. We could not capture these shorter delays (although they did occur frequently) due to the original description of the Delay Tool. In addition, we would not have been able to report data on the impact on LOS and costs, as these are attributed to whole days in the hospital. However, if we had been able to differentiate shorter delays, this would bias our results to show a greater percentage of delays over smaller increments of time. Generalizability is an issue, given that this was a single‐center study. This study sample included over 80 different attending physicians participating in community pediatrician, subspecialty, and hospitalist practice groups. However, the patient population at PCMC is similar to other medium and large children's hospitals in the United States. The month observed may not reflect the entire year of hospitalizationsthere may be seasonal variations with delays depending on the volume and type of illness seen. The study was conducted in August, when there are newer house staff present. However, physician responsibility, which was the largest source of delays in our study, had little attribution to house staff. Most of the decisions were those of attending physicians, which would largely be unaffected by the time of year of the study. Finally, we were unable to assess the safety of the potential earlier discharge, as this was an observational study. However, in any future intervention studies examining processes to discharge patients sooner, measures of safety to the patient are a necessity. Finally, given the potential of ongoing admission, even on the date of discharge of our most fragile patients, this approach to discovering causes of delay may not apply to this important group, which is responsible for significant and growing resource utilization.

Despite these limitations, our findings demonstrate that in an era of children staying in the hospital less, and more medically‐complex children being admitted,10 a substantial number of children who are hospitalized at a children's hospital may have been discharged sooner. The majority of these decisions were directly related to physician responsibility. As consumers, providers, and hospitals work together to develop quality measures that are reflective of inpatient pediatric care, the Delay Tool may be able to highlight 2 aims of quality (ie, timeliness and efficiency of care) that could be used to assess the impact of interventions designed to safely discharge patients sooner. Interventions such as audit‐feedback,18 clinical guideline deployment,19 and hospitalist systems of care4 continue to hold the promise of earlier discharge; however, tools designed to measure inappropriate use of hospital days should be employed to demonstrate their effectiveness. Our study demonstrates ongoing waste in children's hospitals.

Conclusions

Almost 1 out of 4 patients in this 1‐month period could have been discharged sooner than they were. The impact of delays on costs and LOS are substantial and should provide strong incentives to develop effective interventions. Such interventions will need to address variations in physician criteria for discharge, more efficient discharge planning, and timely scheduling of consultation and diagnostic testing.

Inpatient pediatrics is undergoing a paradigm shift in at least 3 ways. First, more children with chronic disease are being cared for in the hospital over time.1 Second, previous inpatient conditions are treated at home with advancing technology such as peripherally‐inserted catheters.2 Third, there are new areas of growing specialization, such as hospital medicine, in which the practitioners deliver more efficient care.3, 4

Nationwide, there is increasing pressure to improve inpatient quality of care. The Institute of Medicine defines 6 aims for improvement, including timeliness (reducing waits and sometimes harmful delays for both those who receive and those who give care) and efficiency of care (avoiding waste, including waste of equipment, supplies, ideas, and energy).5 Reducing unnecessary stays in the hospital is a potential quality measure that hospitals may use to address the timeliness and efficiency of care delivered to hospitalized children.

Delays in discharge have been used as markers of unnecessary stays in the hospital for inpatient adult and pediatric care,6, 7 but these are limited to inpatient systems from almost 20 years ago. Current reasons why patients are delayed from discharge, if at all, are not well described. We undertook this study to describe delays in hospital discharges at a tertiary‐care children's hospital in terms of number of patients, length of days of delay, and type of delay. In addition, we sought to characterize the impact of discharge delays on overall length of stay (LOS) and costs.

Methods

Results

During the 31 days of the study, 171 patients occupied hospital beds an average of 7.3 days on the 2 inpatient medical teams, for a total of 911 inpatient days. Seven patients were admitted prior to August 1; 6 of these were discharged during the month of August and 1 stayed through the entire month and was discharged in September. Three additional patients were admitted in August and discharged in September. There were 6 readmissions during the month of August, and 1 patient was excluded from the study because of lack of sufficient information. All patients with delays were able to be classified according to the Delay Tool taxonomy. Interrater reliability for the 2 study physicians was 98%.

The characteristics between the patients who did and did not experience a delay in discharge are shown in Table 1. Thirty‐nine of 171 patients (22.8%), experienced at least 1 delay day. Eighteen of 39 patients had only 1 delay day (46.2%) and 11 patients experienced 2 delayed days (28.2%) (Figure 1). The average length of delay was 2.1 days.

Figure 1

Delays attributed to physician responsibility accounted for 42.3% (16.5/39) of patient delays (conservative management or clinical decision‐making), with discharge planning delays accounting for 21.8% (family‐related, patient‐related, and hospital‐related problems), consultation for 14.1% (delay in obtaining or lack of follow‐up), test scheduling for 12.8%, and obtaining test results for 5.1% (ordering and weekend scheduling). There were no primary delays due to surgery, education and research, or unavailability of outside resources such as a skilled nursing bed. Four patients had a single additional secondary cause of delay assigned to them, related to physician responsibility, consultation, surgery and test scheduling; these were split, attributing 0.5 patients to each delay type (thus, the 17/39 patients delayed for physician responsibility was analyzed as 16.5/39) (Table 2).

There were 82 delay‐related hospital days of 911 total inpatient days on the 2 medical teams for August 2004 (9%). More than $170,000 in excess costs was incurred due to delay days from a total of approximately 1.9 million dollars in patient costs for the month (8.9%).

Discussion

This study finds that discharge delays in a tertiary care children's hospital are common; almost 1 in 4 patients experienced a medically unnecessary excess hospital stay of at least 1 day. The average length of a delay was 2.1 days, and overall, delays consumed 9% of pediatric hospital days and 8.9% of total costs. The most common reason for a delay was related to physician clinical care, including excessively conservative management and variability in clinical decision‐making.

Our study results are similar to the other 2 published studies that use the Delay Tool. In the adult and pediatric studies, between 10% and 30% of patients experienced a delay in discharge, with the average length of delay between 2.9 and 3 days.6, 7 Although both studies were conducted at teaching hospitals, what is particularly interesting is that they were conducted almost 20 years ago. During this period of time, there has been a shift in the inpatient pediatric patient population. In recent years, children who are cared for in the hospital have more chronic illnesses.1 In addition, there has been a shift in the types of conditions that may be cared for at home and those that now require inpatient stay.2 Despite this, delays continue at a similar proportion, but the cause of delays have shifted from scheduling and consultation to physician responsibility.

There is another tool in the literature which is more widely used, the Pediatric Appropriateness Evaluation Protocol (PAEP), which is based on the Appropriateness Evaluation Protocol for adults.1417 This tool is used to determine the appropriateness of ongoing hospitalization, not the cause of delay if ongoing hospitalization is inappropriate. The 3 areas that are evaluated (medical services, nursing and ancillary services, and patient's condition) have objective criteria that dictate if the hospitalization is appropriate or not (eg, parenteral (intravenous) therapy for at least 8 hours on that day, under nursing and ancillary services). The PAEP may be less sensitive given today's healthcare resource utilization climate. Many clinicians and families would agree that insertion of a peripheral central catheter is an acceptable form of outpatient treatment for many pediatric conditions. In conjunction with the Delay Tool, the PAEP could be used to determine if a delay occurred, then the Delay Tool used to categorize the cause of the delay. We choose to use expert clinician judgment to determine if a delay had occurred. We were more interested in why patients who are admitted (appropriately or inappropriately) cannot be discharged sooner, thus allowing for future intervention studies targeted to impact delays in discharge, as elucidated in this study. The Delay Tool specifically allowed us to categorize the reasons for delays. Given that the average LOS for patients in the nondelayed group was over 4 days, despite not using a tool such as the PAEP, we believe that these were likely to be appropriate admissions.

A recent study reported the first use of the Medical Care Appropriateness Protocol (MCAP) in a tertiary‐care children's hospital. The authors used the MCAP to determine the impact of an intervention on reducing inappropriate hospitals days for children. This tool is similarly labor‐intensive to the Delay Tool. Interestingly, this Canadian study found a high rate of inappropriate hospital days (47%), which may be in part attributable to a different outcome measurement tool and/or a different health care system.18

There are several limitations to our study that deserve mention. The Delay Tool requires clinician judgment regarding whether or not there was a delay in discharge for that day. We may have introduced some bias in our study, as hospitalist investigators assigned the delay and blinding to the attending physician specialty of record was not feasible. However, our results are similar to the other 2 published studies that have used this tool, and we specifically chose not to analyze or report results in terms of hospitalist and nonhospitalist attending physicians. The Delay Tool is not designed to differentiate shorter delays in terms of hours instead of days (eg, due to the inability for the patient to get a ride home). Shorter delays may be of particular importance depending on the occupancy rate of the hospital, the demand for beds, and other patient and hospital factors. We could not capture these shorter delays (although they did occur frequently) due to the original description of the Delay Tool. In addition, we would not have been able to report data on the impact on LOS and costs, as these are attributed to whole days in the hospital. However, if we had been able to differentiate shorter delays, this would bias our results to show a greater percentage of delays over smaller increments of time. Generalizability is an issue, given that this was a single‐center study. This study sample included over 80 different attending physicians participating in community pediatrician, subspecialty, and hospitalist practice groups. However, the patient population at PCMC is similar to other medium and large children's hospitals in the United States. The month observed may not reflect the entire year of hospitalizationsthere may be seasonal variations with delays depending on the volume and type of illness seen. The study was conducted in August, when there are newer house staff present. However, physician responsibility, which was the largest source of delays in our study, had little attribution to house staff. Most of the decisions were those of attending physicians, which would largely be unaffected by the time of year of the study. Finally, we were unable to assess the safety of the potential earlier discharge, as this was an observational study. However, in any future intervention studies examining processes to discharge patients sooner, measures of safety to the patient are a necessity. Finally, given the potential of ongoing admission, even on the date of discharge of our most fragile patients, this approach to discovering causes of delay may not apply to this important group, which is responsible for significant and growing resource utilization.

Despite these limitations, our findings demonstrate that in an era of children staying in the hospital less, and more medically‐complex children being admitted,10 a substantial number of children who are hospitalized at a children's hospital may have been discharged sooner. The majority of these decisions were directly related to physician responsibility. As consumers, providers, and hospitals work together to develop quality measures that are reflective of inpatient pediatric care, the Delay Tool may be able to highlight 2 aims of quality (ie, timeliness and efficiency of care) that could be used to assess the impact of interventions designed to safely discharge patients sooner. Interventions such as audit‐feedback,18 clinical guideline deployment,19 and hospitalist systems of care4 continue to hold the promise of earlier discharge; however, tools designed to measure inappropriate use of hospital days should be employed to demonstrate their effectiveness. Our study demonstrates ongoing waste in children's hospitals.

Conclusions

Almost 1 out of 4 patients in this 1‐month period could have been discharged sooner than they were. The impact of delays on costs and LOS are substantial and should provide strong incentives to develop effective interventions. Such interventions will need to address variations in physician criteria for discharge, more efficient discharge planning, and timely scheduling of consultation and diagnostic testing.

Aspiration pneumonia is the most common cause of death in children with severe neurological impairment (NI).13 For several reasons (eg, improved survival of extremely low‐birth‐weight infants, technological advances, and changing societal attitudes), the number of children with severe NI is increasing. Many children with severe NI have dysfunctional swallowing and gastroesophageal reflux disease (GERD).46 This combination places them at high risk for recurrent aspiration that, in turn, leads to aspiration pneumonia, repeated hospitalization, respiratory failure, compromised quality of life, and death.7, 8

The most common treatment approach for the combination of dysfunctional swallowing and GERD is surgical fundoplication with a gastrostomy feeding tube. Fundoplication is the third most common procedure performed in children by pediatric surgeons in the United States.9 Fifty percent of the children who receive a fundoplication have neurological impairment.10, 11 The goals of the surgery to treat GERD unresponsive to medical management are to reduce the risk of aspiration pneumonia, improve nutritional status, and improve the quality of life of the children and their families. However, few prospective longitudinal studies have determined whether the quality of life of the children or their caregivers actually improves over time.

The importance of caregiver and child quality of life is increasingly recognized as a critical outcome of any intervention in this population.12, 13 Previous studies of fundoplication in children with NI, GERD, and dysfunctional swallowing reported surgical mortality rates between 1% and 3% and other complications between 4% and 39%, reflecting the medical fragility of these children.5, 1418 Some of these studies had longitudinal follow‐up and reported long‐term data. Recurrence of symptoms was reported in up to 56% of patients, recurrence of AP in up to 39%, further surgical procedures in up to 19%, and mortality in up to 17%.14, 1921 Few case series of children with neurological impairment who have had a fundoplication have addressed child and caregiver quality of life following either a fundoplication or placement of a feeding tube.2224 In their study of 16 patients who had a fundoplication and gastrostomy tube placed, Tawfik et al. found improvements in children's happiness, ease of giving medicines, and time to devote to other children. Sullivan et al. found improvement in caregiver quality of life following placement of a gastrostomy tube in a child; however, they did not specifically identify those children who had been treated with a fundoplication. In their retrospective study, O'Neill et al. found improved child and caregiver quality of life following a fundoplication. Collectively, these studies have found that parents report improvement of both their own and their child's quality of life after either intervention. However, not having baseline measurements, not controlling for degree of functional impairment of the children, small sample sizes, and large loss to follow‐up limit the utility of these studies. In this ongoing, long‐term prospective longitudinal study, we report the initial impact of a fundoplication on the quality of life of both children and their caregivers.

The primary objective of this study was to determine change over time in the quality of life of children with neurological impairment who received surgical treatment of their GERD and of the caregivers of these children, controlling for the degree of functional impairment of the children. We hypothesized that child and caregiver quality of life would both improve after primary fundoplication and gastrostomy tube placement. Secondary objectives included describing rates of complications in this population.

METHODS

RESULTS

Sixty‐three children met eligibility criteria. Forty‐four families (70%) initially agreed to participate in the study and completed the baseline questionnaires (see Fig. 1). The mean age of the children was 2.2 years. Twenty‐six parents of children completed the 1‐month postfundoplication quality‐of‐life questionnaires. Thirteen patients were lost to follow‐up, 5 of whom had not reached the 1‐month postfundoplication time point. The median WeeFIM (functional status) score of the whole group was 31.2 (95% confidence interval [CI] 11‐71) compared with a childhood matched‐age norm of 83 (95% CI 60‐110), P = .001. WeeFIM scores did not change significantly from baseline to 1 month, P = .98 (Kruskall‐Wallis test).

Figure 1

Data for the 13 parents and children (30%) who gave baseline data but were subsequently lost to follow‐up are shown in Table 1. Reasons for loss to follow‐up were caregiver reporting being too busy to fill out the questionnaires (n = 8) and no reason stated (n = 5).

In addition to the diagnosis of GERD, clinical indications for fundoplication were vomiting (55%), feeding‐related issues (47%), and failure to thrive (39%). Diagnosis of GERD was confirmed for 41 of 44 patients77% had an abnormal upper GI, 26% an abnormal pH probe, 14% an abnormal endoscopy, and 24% an abnormal swallow study. The remaining 3 had obvious clinical symptoms for GERD and did not require further testing according to their attending surgeon (2 with witnessed vomiting leading to aspiration and 1 who was exclusively gastrostomy‐fed and was witnessed having feeds coming from the tracheostomy). Various medications had been tried and were considered to have failed in these patients: 39% had been treated with acid‐suppressive agents; 80% with acid blocking agents; and 61% with prokinetic agents. Fourteen patients (32%) had cerebral palsy, and 14 (32%) had a brain or spinal cord abnormality (see Table 2).

Thirty‐four children underwent a laparoscopic Nissen fundoplication, and 10 had an open Nissen fundoplication. All had gastrostomy tubes placed or replaced at the time of surgery.

Analysis of the mean bodily pain scores from the CHQ‐PF50 revealed that bodily pain of patients in the study cohort had improved from baseline after 1 month of follow‐up (mean score at baseline, 32.8; after 1 month of follow‐up, 47.5; P = .01), after adjusting for functional status. However, these mean bodily pain scores were significantly lower than those of children with cerebral palsy (mean score, 73.9; P < .001).34, 35 After adjusting for functional status, scores were improved for role/social‐physical limitations (mean baseline score, 30.6; 1‐month follow‐up score, 56.6; P = .01), mental health (mean baseline score, 62.7; 1‐month follow‐up score, 70.6; P = .01), family limitation of activities (mean baseline score, 43.3; 1‐month follow‐up score, 55.1; P = .03), and parental time (mean baseline score, 43.0; 1‐month follow‐up score, 55.3; P = .03). Scores were unchanged for physical function, global health, general health perception, physical summary, role/social‐emotional, mental health, self‐esteem, and psychological summary (see Table 3).

Analysis of the SF‐36 of the parents of these children revealed mean scores significantly lower than those in general U.S. population for all quality‐of‐life domains except physical function (see Table 4). Many baseline domain scores were similar to those of adults with clinical depression. The only domain that showed improvement in quality of life of the caregivers over the 1‐month follow‐up period was vitality (mean baseline score, 41.3; 1‐month follow‐up score, 48.2; P = .001).

Total stress as measured by the PSI/SF mean was 79.1 at baseline and 77.6 1 month after fundoplication (P = .54). This was significantly higher stress than the parental norm of 71.0 (P = .01). One in 4 parents expressed clinically significant levels of stress (scores > 90, 90th percentile).

Patients suffered the following complications in the month after fundoplication. Eight children had at least 1 subsequent emergency department visit related to a complication of the gastrostomy tube (8 visits), to respiratory distress (1 visit), or tovomiting (1 visit). Seven children had a subsequent admission to the hospital related to a complication of the gastrostomy tube (4 admissions), complication of surgery (2 admissions), or aspiration pneumonia (1 admission). None of the children had a repeat fundoplication or subsequently underwent placement of a gastrojejunal feeding tube. One patient died. She was 10 months old when she died, which was 3 weeks after she had received a fundoplication. She had obstructive hydrocephalus, cortical blindness, and developmental delay, and respiratory arrest and subsequent tonsillar herniation led to her death.

DISCUSSION

Parents of children with neurological impairment and GERD who underwent their first fundoplication reported improved quality of life of their children in the domains of bodily pain, role/social‐physical limitations, mental health, family limitation of activities, and parental time over the first month after surgery, when controlling for the children's degree of functional impairment. The only significant similar improvement in the parent self‐reported quality of life was in the domain of vitality.

This study had several limitations. Loss to follow‐up may have led to a bias reflecting the phenomenon that patients who have poorer quality of life are less likely to report this, or even to be able to participate in the follow‐up component of a study like this. In survival analyses, this incomplete follow‐up of patients is called informative dropout and may be minimized by applying a statistical technique that accounts for this, using the Q‐TWiST.45 However, our current study design and analysis plan precluded using this methodology. As shown in Table 1, we did not find any differences between those patients who stayed in the study and those who dropped out. Also, we were able to contact most parents who reported being too busy to fill out the surveys. Patient heterogeneity is also a concern: Table 2 shows the wide array of diagnoses responsible for the children's neurological impairment. However, we used a standardized functional status measure to ensure we were analyzing similarly disabled patients. Also, the standard deviation of the mean WeeFIM score was small, implying little variability in the study cohort. Our study analyzed data from a single center, which reflects care in the western United States. However, our hospital is similar to other medium and large children's hospitals and our patient population similar to others that perform fundoplication for children with neurological impairment.46 We believe our findings are generalizable to other surgical centers that perform a similar volume of fundoplications in such children with NI.

Our study findings are similar to those reported by O'Neill et al., whose study found that parents reported improved quality of life of their children in ease of feedings, physical comfort during feeding, and ability of the child to enjoy life.23 The CHQPF50 does not specifically ask about feeding, but we did find similar improvement in the domain of role/social‐physical limitations. O'Neill et al. also found that after the children in their study received a fundoplication, caregivers reported their own quality of life improved in the areas of being able to spend more time caring for their child's needs, which is similar to our findings of fewer family limitations of activities and more parental time. Our findings were somewhat dissimilar to the O'Neill et al. study, as parents in their study found several additional areas of improvement in caregiver own quality of life. One explanation for the differing results may be differences in the populations studied. Parents in our study had SF‐36 scores for general health, vitality, and social functioning that were similar to those of adults with depression,47 whereas parents in the O'Neill study did not. Although the O'Neill et al. study was the first to examine these critical quality‐of‐life outcomes for children with NI who have received fundoplication, it had several methodological limitations. We have had the opportunity to build on the work of O'Neill et al. and in a prospective study to capture standardized baseline data (therefore not subject to recall bias, as was likely in the O'Neill et al. study) and collect long‐term data on this population. We also controlled for functional status, which did not improve over the 1 month and by itself could be responsible for the already poor caregiver quality of life. Some aspects of the children's care did improve, but perhaps not enough to overcome the severe disabilities the children and their caregivers live with on a daily basis. We found some evidence to support that the parents' PSI/SF scores were similar to those of parents of children with heart disease, other enterally fed children, and children with traumatic brain injury (who make up between 1 in 3 and 1 in 5 parents with severe stress).39, 41, 43 Future interventions should address the stress and quality of life of these caregivers, especially if surgery does not improve caregiver quality of life or decrease stress.

Contrary to an emerging body of literature in pediatrics that describes a positive correlation between the health of children with chronic illnesses and their caregivers' quality of life,12, 42 we did not find large immediate improvements in caregiver quality of life and decrease in stress as their children's quality of life improved. This may be related to the number of parents in our sample being too small to detect such changes or that changes in longer‐term (greater than 6 months or 1 year) quality of life not being reflected by short‐term assessment. Caregiver and child quality of life following fundoplication needs to be studied over the long term (eg, over many years). We are continuing to follow these patients and their families and will repeat the quality‐of‐life measures 6 and 12 months after fundoplication and report these findings.

Additional studies of treatments for neurologically impaired children with GERD are needed. Randomized trials of alternatives to fundoplication such as gastrojejunal feeding tubes have been proposed, with which we strongly agree.46, 48 We believe that any randomized, controlled trial of children with neurological impairment and GERD must measure child and caregiver quality of life and functional status outcomes. 0

Aspiration pneumonia is the most common cause of death in children with severe neurological impairment (NI).13 For several reasons (eg, improved survival of extremely low‐birth‐weight infants, technological advances, and changing societal attitudes), the number of children with severe NI is increasing. Many children with severe NI have dysfunctional swallowing and gastroesophageal reflux disease (GERD).46 This combination places them at high risk for recurrent aspiration that, in turn, leads to aspiration pneumonia, repeated hospitalization, respiratory failure, compromised quality of life, and death.7, 8

The most common treatment approach for the combination of dysfunctional swallowing and GERD is surgical fundoplication with a gastrostomy feeding tube. Fundoplication is the third most common procedure performed in children by pediatric surgeons in the United States.9 Fifty percent of the children who receive a fundoplication have neurological impairment.10, 11 The goals of the surgery to treat GERD unresponsive to medical management are to reduce the risk of aspiration pneumonia, improve nutritional status, and improve the quality of life of the children and their families. However, few prospective longitudinal studies have determined whether the quality of life of the children or their caregivers actually improves over time.

The importance of caregiver and child quality of life is increasingly recognized as a critical outcome of any intervention in this population.12, 13 Previous studies of fundoplication in children with NI, GERD, and dysfunctional swallowing reported surgical mortality rates between 1% and 3% and other complications between 4% and 39%, reflecting the medical fragility of these children.5, 1418 Some of these studies had longitudinal follow‐up and reported long‐term data. Recurrence of symptoms was reported in up to 56% of patients, recurrence of AP in up to 39%, further surgical procedures in up to 19%, and mortality in up to 17%.14, 1921 Few case series of children with neurological impairment who have had a fundoplication have addressed child and caregiver quality of life following either a fundoplication or placement of a feeding tube.2224 In their study of 16 patients who had a fundoplication and gastrostomy tube placed, Tawfik et al. found improvements in children's happiness, ease of giving medicines, and time to devote to other children. Sullivan et al. found improvement in caregiver quality of life following placement of a gastrostomy tube in a child; however, they did not specifically identify those children who had been treated with a fundoplication. In their retrospective study, O'Neill et al. found improved child and caregiver quality of life following a fundoplication. Collectively, these studies have found that parents report improvement of both their own and their child's quality of life after either intervention. However, not having baseline measurements, not controlling for degree of functional impairment of the children, small sample sizes, and large loss to follow‐up limit the utility of these studies. In this ongoing, long‐term prospective longitudinal study, we report the initial impact of a fundoplication on the quality of life of both children and their caregivers.

The primary objective of this study was to determine change over time in the quality of life of children with neurological impairment who received surgical treatment of their GERD and of the caregivers of these children, controlling for the degree of functional impairment of the children. We hypothesized that child and caregiver quality of life would both improve after primary fundoplication and gastrostomy tube placement. Secondary objectives included describing rates of complications in this population.

METHODS

RESULTS

Sixty‐three children met eligibility criteria. Forty‐four families (70%) initially agreed to participate in the study and completed the baseline questionnaires (see Fig. 1). The mean age of the children was 2.2 years. Twenty‐six parents of children completed the 1‐month postfundoplication quality‐of‐life questionnaires. Thirteen patients were lost to follow‐up, 5 of whom had not reached the 1‐month postfundoplication time point. The median WeeFIM (functional status) score of the whole group was 31.2 (95% confidence interval [CI] 11‐71) compared with a childhood matched‐age norm of 83 (95% CI 60‐110), P = .001. WeeFIM scores did not change significantly from baseline to 1 month, P = .98 (Kruskall‐Wallis test).

Figure 1

Data for the 13 parents and children (30%) who gave baseline data but were subsequently lost to follow‐up are shown in Table 1. Reasons for loss to follow‐up were caregiver reporting being too busy to fill out the questionnaires (n = 8) and no reason stated (n = 5).

In addition to the diagnosis of GERD, clinical indications for fundoplication were vomiting (55%), feeding‐related issues (47%), and failure to thrive (39%). Diagnosis of GERD was confirmed for 41 of 44 patients77% had an abnormal upper GI, 26% an abnormal pH probe, 14% an abnormal endoscopy, and 24% an abnormal swallow study. The remaining 3 had obvious clinical symptoms for GERD and did not require further testing according to their attending surgeon (2 with witnessed vomiting leading to aspiration and 1 who was exclusively gastrostomy‐fed and was witnessed having feeds coming from the tracheostomy). Various medications had been tried and were considered to have failed in these patients: 39% had been treated with acid‐suppressive agents; 80% with acid blocking agents; and 61% with prokinetic agents. Fourteen patients (32%) had cerebral palsy, and 14 (32%) had a brain or spinal cord abnormality (see Table 2).

Thirty‐four children underwent a laparoscopic Nissen fundoplication, and 10 had an open Nissen fundoplication. All had gastrostomy tubes placed or replaced at the time of surgery.

Analysis of the mean bodily pain scores from the CHQ‐PF50 revealed that bodily pain of patients in the study cohort had improved from baseline after 1 month of follow‐up (mean score at baseline, 32.8; after 1 month of follow‐up, 47.5; P = .01), after adjusting for functional status. However, these mean bodily pain scores were significantly lower than those of children with cerebral palsy (mean score, 73.9; P < .001).34, 35 After adjusting for functional status, scores were improved for role/social‐physical limitations (mean baseline score, 30.6; 1‐month follow‐up score, 56.6; P = .01), mental health (mean baseline score, 62.7; 1‐month follow‐up score, 70.6; P = .01), family limitation of activities (mean baseline score, 43.3; 1‐month follow‐up score, 55.1; P = .03), and parental time (mean baseline score, 43.0; 1‐month follow‐up score, 55.3; P = .03). Scores were unchanged for physical function, global health, general health perception, physical summary, role/social‐emotional, mental health, self‐esteem, and psychological summary (see Table 3).

Analysis of the SF‐36 of the parents of these children revealed mean scores significantly lower than those in general U.S. population for all quality‐of‐life domains except physical function (see Table 4). Many baseline domain scores were similar to those of adults with clinical depression. The only domain that showed improvement in quality of life of the caregivers over the 1‐month follow‐up period was vitality (mean baseline score, 41.3; 1‐month follow‐up score, 48.2; P = .001).

Total stress as measured by the PSI/SF mean was 79.1 at baseline and 77.6 1 month after fundoplication (P = .54). This was significantly higher stress than the parental norm of 71.0 (P = .01). One in 4 parents expressed clinically significant levels of stress (scores > 90, 90th percentile).

Patients suffered the following complications in the month after fundoplication. Eight children had at least 1 subsequent emergency department visit related to a complication of the gastrostomy tube (8 visits), to respiratory distress (1 visit), or tovomiting (1 visit). Seven children had a subsequent admission to the hospital related to a complication of the gastrostomy tube (4 admissions), complication of surgery (2 admissions), or aspiration pneumonia (1 admission). None of the children had a repeat fundoplication or subsequently underwent placement of a gastrojejunal feeding tube. One patient died. She was 10 months old when she died, which was 3 weeks after she had received a fundoplication. She had obstructive hydrocephalus, cortical blindness, and developmental delay, and respiratory arrest and subsequent tonsillar herniation led to her death.

DISCUSSION

Parents of children with neurological impairment and GERD who underwent their first fundoplication reported improved quality of life of their children in the domains of bodily pain, role/social‐physical limitations, mental health, family limitation of activities, and parental time over the first month after surgery, when controlling for the children's degree of functional impairment. The only significant similar improvement in the parent self‐reported quality of life was in the domain of vitality.

This study had several limitations. Loss to follow‐up may have led to a bias reflecting the phenomenon that patients who have poorer quality of life are less likely to report this, or even to be able to participate in the follow‐up component of a study like this. In survival analyses, this incomplete follow‐up of patients is called informative dropout and may be minimized by applying a statistical technique that accounts for this, using the Q‐TWiST.45 However, our current study design and analysis plan precluded using this methodology. As shown in Table 1, we did not find any differences between those patients who stayed in the study and those who dropped out. Also, we were able to contact most parents who reported being too busy to fill out the surveys. Patient heterogeneity is also a concern: Table 2 shows the wide array of diagnoses responsible for the children's neurological impairment. However, we used a standardized functional status measure to ensure we were analyzing similarly disabled patients. Also, the standard deviation of the mean WeeFIM score was small, implying little variability in the study cohort. Our study analyzed data from a single center, which reflects care in the western United States. However, our hospital is similar to other medium and large children's hospitals and our patient population similar to others that perform fundoplication for children with neurological impairment.46 We believe our findings are generalizable to other surgical centers that perform a similar volume of fundoplications in such children with NI.

Our study findings are similar to those reported by O'Neill et al., whose study found that parents reported improved quality of life of their children in ease of feedings, physical comfort during feeding, and ability of the child to enjoy life.23 The CHQPF50 does not specifically ask about feeding, but we did find similar improvement in the domain of role/social‐physical limitations. O'Neill et al. also found that after the children in their study received a fundoplication, caregivers reported their own quality of life improved in the areas of being able to spend more time caring for their child's needs, which is similar to our findings of fewer family limitations of activities and more parental time. Our findings were somewhat dissimilar to the O'Neill et al. study, as parents in their study found several additional areas of improvement in caregiver own quality of life. One explanation for the differing results may be differences in the populations studied. Parents in our study had SF‐36 scores for general health, vitality, and social functioning that were similar to those of adults with depression,47 whereas parents in the O'Neill study did not. Although the O'Neill et al. study was the first to examine these critical quality‐of‐life outcomes for children with NI who have received fundoplication, it had several methodological limitations. We have had the opportunity to build on the work of O'Neill et al. and in a prospective study to capture standardized baseline data (therefore not subject to recall bias, as was likely in the O'Neill et al. study) and collect long‐term data on this population. We also controlled for functional status, which did not improve over the 1 month and by itself could be responsible for the already poor caregiver quality of life. Some aspects of the children's care did improve, but perhaps not enough to overcome the severe disabilities the children and their caregivers live with on a daily basis. We found some evidence to support that the parents' PSI/SF scores were similar to those of parents of children with heart disease, other enterally fed children, and children with traumatic brain injury (who make up between 1 in 3 and 1 in 5 parents with severe stress).39, 41, 43 Future interventions should address the stress and quality of life of these caregivers, especially if surgery does not improve caregiver quality of life or decrease stress.

Contrary to an emerging body of literature in pediatrics that describes a positive correlation between the health of children with chronic illnesses and their caregivers' quality of life,12, 42 we did not find large immediate improvements in caregiver quality of life and decrease in stress as their children's quality of life improved. This may be related to the number of parents in our sample being too small to detect such changes or that changes in longer‐term (greater than 6 months or 1 year) quality of life not being reflected by short‐term assessment. Caregiver and child quality of life following fundoplication needs to be studied over the long term (eg, over many years). We are continuing to follow these patients and their families and will repeat the quality‐of‐life measures 6 and 12 months after fundoplication and report these findings.

Additional studies of treatments for neurologically impaired children with GERD are needed. Randomized trials of alternatives to fundoplication such as gastrojejunal feeding tubes have been proposed, with which we strongly agree.46, 48 We believe that any randomized, controlled trial of children with neurological impairment and GERD must measure child and caregiver quality of life and functional status outcomes. 0

Aspiration pneumonia is the most common cause of death in children with severe neurological impairment (NI).13 For several reasons (eg, improved survival of extremely low‐birth‐weight infants, technological advances, and changing societal attitudes), the number of children with severe NI is increasing. Many children with severe NI have dysfunctional swallowing and gastroesophageal reflux disease (GERD).46 This combination places them at high risk for recurrent aspiration that, in turn, leads to aspiration pneumonia, repeated hospitalization, respiratory failure, compromised quality of life, and death.7, 8

The most common treatment approach for the combination of dysfunctional swallowing and GERD is surgical fundoplication with a gastrostomy feeding tube. Fundoplication is the third most common procedure performed in children by pediatric surgeons in the United States.9 Fifty percent of the children who receive a fundoplication have neurological impairment.10, 11 The goals of the surgery to treat GERD unresponsive to medical management are to reduce the risk of aspiration pneumonia, improve nutritional status, and improve the quality of life of the children and their families. However, few prospective longitudinal studies have determined whether the quality of life of the children or their caregivers actually improves over time.

The importance of caregiver and child quality of life is increasingly recognized as a critical outcome of any intervention in this population.12, 13 Previous studies of fundoplication in children with NI, GERD, and dysfunctional swallowing reported surgical mortality rates between 1% and 3% and other complications between 4% and 39%, reflecting the medical fragility of these children.5, 1418 Some of these studies had longitudinal follow‐up and reported long‐term data. Recurrence of symptoms was reported in up to 56% of patients, recurrence of AP in up to 39%, further surgical procedures in up to 19%, and mortality in up to 17%.14, 1921 Few case series of children with neurological impairment who have had a fundoplication have addressed child and caregiver quality of life following either a fundoplication or placement of a feeding tube.2224 In their study of 16 patients who had a fundoplication and gastrostomy tube placed, Tawfik et al. found improvements in children's happiness, ease of giving medicines, and time to devote to other children. Sullivan et al. found improvement in caregiver quality of life following placement of a gastrostomy tube in a child; however, they did not specifically identify those children who had been treated with a fundoplication. In their retrospective study, O'Neill et al. found improved child and caregiver quality of life following a fundoplication. Collectively, these studies have found that parents report improvement of both their own and their child's quality of life after either intervention. However, not having baseline measurements, not controlling for degree of functional impairment of the children, small sample sizes, and large loss to follow‐up limit the utility of these studies. In this ongoing, long‐term prospective longitudinal study, we report the initial impact of a fundoplication on the quality of life of both children and their caregivers.

The primary objective of this study was to determine change over time in the quality of life of children with neurological impairment who received surgical treatment of their GERD and of the caregivers of these children, controlling for the degree of functional impairment of the children. We hypothesized that child and caregiver quality of life would both improve after primary fundoplication and gastrostomy tube placement. Secondary objectives included describing rates of complications in this population.

METHODS

RESULTS

Sixty‐three children met eligibility criteria. Forty‐four families (70%) initially agreed to participate in the study and completed the baseline questionnaires (see Fig. 1). The mean age of the children was 2.2 years. Twenty‐six parents of children completed the 1‐month postfundoplication quality‐of‐life questionnaires. Thirteen patients were lost to follow‐up, 5 of whom had not reached the 1‐month postfundoplication time point. The median WeeFIM (functional status) score of the whole group was 31.2 (95% confidence interval [CI] 11‐71) compared with a childhood matched‐age norm of 83 (95% CI 60‐110), P = .001. WeeFIM scores did not change significantly from baseline to 1 month, P = .98 (Kruskall‐Wallis test).

Figure 1

Data for the 13 parents and children (30%) who gave baseline data but were subsequently lost to follow‐up are shown in Table 1. Reasons for loss to follow‐up were caregiver reporting being too busy to fill out the questionnaires (n = 8) and no reason stated (n = 5).

In addition to the diagnosis of GERD, clinical indications for fundoplication were vomiting (55%), feeding‐related issues (47%), and failure to thrive (39%). Diagnosis of GERD was confirmed for 41 of 44 patients77% had an abnormal upper GI, 26% an abnormal pH probe, 14% an abnormal endoscopy, and 24% an abnormal swallow study. The remaining 3 had obvious clinical symptoms for GERD and did not require further testing according to their attending surgeon (2 with witnessed vomiting leading to aspiration and 1 who was exclusively gastrostomy‐fed and was witnessed having feeds coming from the tracheostomy). Various medications had been tried and were considered to have failed in these patients: 39% had been treated with acid‐suppressive agents; 80% with acid blocking agents; and 61% with prokinetic agents. Fourteen patients (32%) had cerebral palsy, and 14 (32%) had a brain or spinal cord abnormality (see Table 2).

Thirty‐four children underwent a laparoscopic Nissen fundoplication, and 10 had an open Nissen fundoplication. All had gastrostomy tubes placed or replaced at the time of surgery.

Analysis of the mean bodily pain scores from the CHQ‐PF50 revealed that bodily pain of patients in the study cohort had improved from baseline after 1 month of follow‐up (mean score at baseline, 32.8; after 1 month of follow‐up, 47.5; P = .01), after adjusting for functional status. However, these mean bodily pain scores were significantly lower than those of children with cerebral palsy (mean score, 73.9; P < .001).34, 35 After adjusting for functional status, scores were improved for role/social‐physical limitations (mean baseline score, 30.6; 1‐month follow‐up score, 56.6; P = .01), mental health (mean baseline score, 62.7; 1‐month follow‐up score, 70.6; P = .01), family limitation of activities (mean baseline score, 43.3; 1‐month follow‐up score, 55.1; P = .03), and parental time (mean baseline score, 43.0; 1‐month follow‐up score, 55.3; P = .03). Scores were unchanged for physical function, global health, general health perception, physical summary, role/social‐emotional, mental health, self‐esteem, and psychological summary (see Table 3).

Analysis of the SF‐36 of the parents of these children revealed mean scores significantly lower than those in general U.S. population for all quality‐of‐life domains except physical function (see Table 4). Many baseline domain scores were similar to those of adults with clinical depression. The only domain that showed improvement in quality of life of the caregivers over the 1‐month follow‐up period was vitality (mean baseline score, 41.3; 1‐month follow‐up score, 48.2; P = .001).

Total stress as measured by the PSI/SF mean was 79.1 at baseline and 77.6 1 month after fundoplication (P = .54). This was significantly higher stress than the parental norm of 71.0 (P = .01). One in 4 parents expressed clinically significant levels of stress (scores > 90, 90th percentile).

Patients suffered the following complications in the month after fundoplication. Eight children had at least 1 subsequent emergency department visit related to a complication of the gastrostomy tube (8 visits), to respiratory distress (1 visit), or tovomiting (1 visit). Seven children had a subsequent admission to the hospital related to a complication of the gastrostomy tube (4 admissions), complication of surgery (2 admissions), or aspiration pneumonia (1 admission). None of the children had a repeat fundoplication or subsequently underwent placement of a gastrojejunal feeding tube. One patient died. She was 10 months old when she died, which was 3 weeks after she had received a fundoplication. She had obstructive hydrocephalus, cortical blindness, and developmental delay, and respiratory arrest and subsequent tonsillar herniation led to her death.

DISCUSSION

Parents of children with neurological impairment and GERD who underwent their first fundoplication reported improved quality of life of their children in the domains of bodily pain, role/social‐physical limitations, mental health, family limitation of activities, and parental time over the first month after surgery, when controlling for the children's degree of functional impairment. The only significant similar improvement in the parent self‐reported quality of life was in the domain of vitality.

This study had several limitations. Loss to follow‐up may have led to a bias reflecting the phenomenon that patients who have poorer quality of life are less likely to report this, or even to be able to participate in the follow‐up component of a study like this. In survival analyses, this incomplete follow‐up of patients is called informative dropout and may be minimized by applying a statistical technique that accounts for this, using the Q‐TWiST.45 However, our current study design and analysis plan precluded using this methodology. As shown in Table 1, we did not find any differences between those patients who stayed in the study and those who dropped out. Also, we were able to contact most parents who reported being too busy to fill out the surveys. Patient heterogeneity is also a concern: Table 2 shows the wide array of diagnoses responsible for the children's neurological impairment. However, we used a standardized functional status measure to ensure we were analyzing similarly disabled patients. Also, the standard deviation of the mean WeeFIM score was small, implying little variability in the study cohort. Our study analyzed data from a single center, which reflects care in the western United States. However, our hospital is similar to other medium and large children's hospitals and our patient population similar to others that perform fundoplication for children with neurological impairment.46 We believe our findings are generalizable to other surgical centers that perform a similar volume of fundoplications in such children with NI.

Our study findings are similar to those reported by O'Neill et al., whose study found that parents reported improved quality of life of their children in ease of feedings, physical comfort during feeding, and ability of the child to enjoy life.23 The CHQPF50 does not specifically ask about feeding, but we did find similar improvement in the domain of role/social‐physical limitations. O'Neill et al. also found that after the children in their study received a fundoplication, caregivers reported their own quality of life improved in the areas of being able to spend more time caring for their child's needs, which is similar to our findings of fewer family limitations of activities and more parental time. Our findings were somewhat dissimilar to the O'Neill et al. study, as parents in their study found several additional areas of improvement in caregiver own quality of life. One explanation for the differing results may be differences in the populations studied. Parents in our study had SF‐36 scores for general health, vitality, and social functioning that were similar to those of adults with depression,47 whereas parents in the O'Neill study did not. Although the O'Neill et al. study was the first to examine these critical quality‐of‐life outcomes for children with NI who have received fundoplication, it had several methodological limitations. We have had the opportunity to build on the work of O'Neill et al. and in a prospective study to capture standardized baseline data (therefore not subject to recall bias, as was likely in the O'Neill et al. study) and collect long‐term data on this population. We also controlled for functional status, which did not improve over the 1 month and by itself could be responsible for the already poor caregiver quality of life. Some aspects of the children's care did improve, but perhaps not enough to overcome the severe disabilities the children and their caregivers live with on a daily basis. We found some evidence to support that the parents' PSI/SF scores were similar to those of parents of children with heart disease, other enterally fed children, and children with traumatic brain injury (who make up between 1 in 3 and 1 in 5 parents with severe stress).39, 41, 43 Future interventions should address the stress and quality of life of these caregivers, especially if surgery does not improve caregiver quality of life or decrease stress.

Contrary to an emerging body of literature in pediatrics that describes a positive correlation between the health of children with chronic illnesses and their caregivers' quality of life,12, 42 we did not find large immediate improvements in caregiver quality of life and decrease in stress as their children's quality of life improved. This may be related to the number of parents in our sample being too small to detect such changes or that changes in longer‐term (greater than 6 months or 1 year) quality of life not being reflected by short‐term assessment. Caregiver and child quality of life following fundoplication needs to be studied over the long term (eg, over many years). We are continuing to follow these patients and their families and will repeat the quality‐of‐life measures 6 and 12 months after fundoplication and report these findings.

Additional studies of treatments for neurologically impaired children with GERD are needed. Randomized trials of alternatives to fundoplication such as gastrojejunal feeding tubes have been proposed, with which we strongly agree.46, 48 We believe that any randomized, controlled trial of children with neurological impairment and GERD must measure child and caregiver quality of life and functional status outcomes. 0