Demand and Predictors for Post-Discharge Medical Counseling in Home Care Patients: A Prospective Cohort Study

Rationale Post-discharge care is challenging due to the high rate of adverse events after discharge. However, details regarding post-discharge care requirements remain unclear. Post-discharge medical counseling (PDMC) by telephone service was set-up to investigate its demand and predictors. Methods This prospective study was conducted from April 2011 to March 2012 in a tertiary referral center in northern Taiwan. Patients discharged for home care were recruited and educated via telephone hotline counseling when needed. The patient’s characteristics and call-in details were recorded, and predictors of PDMC use and worsening by red-flag sign were analyzed. Results During the study period, 224 patients were enrolled. The PDMC was used 121 times by 65 patients in an average of 8.6 days after discharge. The red-flag sign was noted in 17 PDMC from 16 patients. Of the PDMC used, 50% (n = 60) were for symptom change and the rest were for post-discharge care problems and issues regarding other administrative services. Predictors of PDMC were underlying malignancy and lower Barthel index (BI). On the other hand, lower BI, higher adjusted Charlson co-morbidity index (CCI), and longer length of hospital stay were associated with PDMC and red-flag sign. Conclusions Demand for PDMC may be as high as 29% in home care patients within 30 days after discharge. PDMC is needed more by patients with malignancy and lower BI. More focus should also be given to those with lower BI, higher CCI, and longer length of hospital stay, as they more frequently have red flag signs.


Introduction
Current global ageing is continuously progressing [1,2] and hospitalization demands increase every year [3]. For in-patient care, hospitalist medicine continues to grow and is now mainstream treatment [4][5][6][7][8]. But even though the hospitalist system reduces hospitalization costs, a major concern is the continuity of patient care, which, when interrupted after discharge, correlates with adverse events and leads to readmission [9,10].
In fact, readmission rates are very high for elderly patients and are reported to be around 20% within one month after discharge [7,11]. Several possible reasons include instability of chronic disease and insufficient communication among physicians [12], yet the probability of post-discharge requirements for help and their classifications are not very clear. Patients with different types of requirements deserve different types of transitional care.
Thus, post-discharge medical counseling (PDMC) using the telephone was set up to investigate the needs for medical help and provide support before the occurrence of readmission, especially for patients cared for at home. Predictors and their association with adverse events after discharge were also analyzed.

Study Subjects
This prospective study was conducted at the National Taiwan University Hospital, a tertiary-care referral center in northern Taiwan. From April 2011 to March 2012, all patients aged .20 years old and admitted to the hospitalist-care ward were consecutively screened. In-patient care was performed by hospitalists. In Taiwan, the hospitalist-care system and primarycare physicians system for in-patient care exist together currently [7]. Patients admitted to one-half of the ward were recruited because another investigation was conducted in the other half. Potential participants from patients who were discharged alive for home care were identified. Other eligibility criteria included having a telephone line at home without communication deficits.
Patients were excluded if they were repeatedly admitted, required subspecialty care, or were enrolled in another postdischarge study. The study was approved by the Research Ethics Committee of National Taiwan University Hospital. All participants provided written informed consent.

Study Protocol
When up for discharge, the enrolled patients and their caregivers were educated regarding the standard care plan and post-discharge monitoring methods based on their illness, such as blood glucose measurement for diabetes mellitus. They were given the contact information of the PDMC that they could call if they needed clinical advice. The telephone hot-line for the PDMC was free and was maintained from 8 a.m. to 8 p.m. every workday and 8 a.m. to 5 p.m. on the weekend.
In the PDMC contact, the case manager who had a nursing certificate took first-line responsibility for responding to the counseling needs, which included health/diet/drug instructions, care skill explanations, and clinic appointments. If the problem was beyond the ability of the case manager, she called for help from a nurse practitioner and then an attending physician until the issue was resolved. The case managers could choose to call an oncall attending physician directly if a red-flag sign was present. Redflag signs (RFS) ( Table 1), indicated a worsening condition and were defined by a consensus in a round table meeting as outlined in a previous study [13]. These RFS consisted of 1) abnormal vital signs like high blood pressure, fever, tachycardia, and dyspnea; 2) abnormal findings related to daily care, including consciousness level, BI score, stool, sputum, urine output, and body weight; and 3) progression of specific findings such as edema, and local skin lesion. If further medical care was required, the patient was referred to the clinic or emergency department because in Taiwan, prescriptions are given only after a face to face visit.

Clinical Characteristics
The patients' clinical characteristics, laboratory data, hospital course, and outcomes were recorded by the case managers. A unified case report form containing a default option for selection was used in order to avoid ambiguous data coding. The primary endpoints were the request for PDMC and the presence of RFS. Other outcomes after discharge included unplanned visits to the emergency department and unplanned readmission.
The Charlson co-morbidity index (CCI) and Barthel index (BI) scores were calculated as in previous studies [14,15]. A BI score .70 was defined as independent performance of activities of daily life, #35 as dependent, and the remainder as intermediate grade.
Underlying malignancy was defined as active cancer without mention of cure or remission. Anemia was defined as hemoglobin ,12 g/dL in males and ,11 g/dL in females. A primary care physician was defined when the patient visited the same doctor three times or more within one year prior to this admission [9]. Artificial tube/catheter included naso-gastric tube, tracheostomy tube, draining tube, Foley catheter, and catheter for dialysis.

Statistical Analysis
Inter-group differences were compared using independent t test for numerical variables and chi-square test for categorical variables. By the stepwise method, multivariate Cox proportional hazard regression was used to identify factors associated with demand for PDMC. A two-sided p,0.05 was considered to be significant. Survival curves were generated using the Kaplan-Meier method and compared using the log-rank test. All analyses were performed using the SPSS (Version 15.0, Chicago, IL).

Results
From April 2011 to March 2012, 1118 patients were admitted from the emergency department to the hospitalist-care ward. Of the 712 patients discharged for home care, 351 eligible patients were invited and 224 finally enrolled (Fig. 1). Within 30 days after discharge, PDMC were requested 121 times from 65 (29%) patients. Among them, RFS was found in 17 events from 16 (7%) patients.
Of the 65 patients who used PDMC, 24 (37%) called more than once and 4 (6%) called more than five times ( Fig. 2A). The PDMC was needed mostly in the first week post-discharge, and then decreased thereafter (Fig. 2B). In terms of the nature of the call-in, the issues identified were problems related to major illnesses in the last admission (n = 45; 37%), new symptoms/problems (n = 15; 13%), and questions unrelated to changes in medical illness (n = 61; 50%), which include 26 for examination/clinic arrangements, 5 for certification, and 14, 11, and 5 for general, drug, and tube/wound instructions, respectively. Advice and explanations were given to all patients. Later, 31 patients (26%) were asked to receive clinical services.
There were 16 patients with 17 PDMC plus RFS, which occurred with a mean of 8.2 days post-discharge. These included 13 events associated with major illness on the last admission and 4 with new symptoms/problems. Nine patients were referred to the ED and seven for clinical management.
Comparing the patients who used the PDMC to those who did not (Table 2), the clinical characteristics were similar except for the poorer renal function, higher BI score at discharge, and less tube/ catheter use in patients who did not use the PDMC. In contrast, patients with PDMC plus RFS had higher adjusted CCI score, longer hospital stay, and lower BI score than those without RFS.
In the post-discharge course (Table 3), patients with PDMC plus RFS had higher rates of emergency department visits and readmissions (94% and 56%, respectively) than patients with PDMC only (48% and 29%, respectively) and those without PDMC. Unexpected death after discharge was similar in all three groups.
Multivariate Cox regression was performed for PDMC within 30 days post-discharge using clinically significant factors (  (Fig. 3).
For the additional resources spent on maintaining a part-time PDMC, all of the time the staff used in the process of enrollment and medical counseling were recorded. Based on staff's regular payment, the approximate costs spent for this service was calculated. Approximately US dollar (US$) 903.50 was used for the recruitment process and US$ 326.90 for the counseling service (exchange rate of US$ 1.00 to NT$ [New Taiwan dollar] 29.37 as of 1st October 2012) ( Table 5). If the costs were divided among the 224 enrolled patients, the cost was US $5.5 per patient for PDMC use in the first month post-discharge.

Discussion
In the present study, the demand for PDMC is high (29%) among home care patients within 30 days post-discharge. Half of the PDMC inquiries were related to symptom management, while the rest were for care skill instructions (24%) and clinic/exam or certification arrangements (26%). With or without RFS, PDMC use was associated with higher emergency department visits and unplanned readmissions. Predictors of PDMC use were underlying malignancy and lower BI score at discharge, whereas higher CCI score, longer hospital stay, and lower BI score were predictive of RFS.
Early after discharge, a patient's condition may dynamically and easily change. One of the reasons is poor transition of care service from the hospital setting to home care [12,16]. Although the demand for PDMC is high, a fourth of them were requests regarding the requirements of administrative services, such as clinic/examination arrangements or certificate applications, while another one-fourth are related to the needs of care instructions such as knowledge/techniques learning. In short, care instructions and administrative services account for 50% of the counseling. For these, pre-discharge training and strengthening the way in which instructions are given can reduce the demand for PDMC and may prevent worsening of conditions. For instance, by improving postdischarge drug adherence, the occurrence of readmission can also be reduced, based on a previous report [17]. Moreover, in-patients cared for by primary care physicians may also decrease the demand for PDMC due to its lower care discontinuity [10].
On the other hand, half of the PDMCs are associated with symptom management. The proportion for symptoms counseling is similar to that of a study from the United States (14.5% vs. 13%) [18]. Among them, RFS is highly associated with post-discharge adverse events and should be considered as warning of worsening medical condition [13]. Predictors for PDMC and RFS can be used to target groups for monitoring when resources are limited. Among the predictors, activities of daily life measured by BI are associated with demands for PDMC with RFS. One possible explanation is that patients with lower BI are more dependent and have more tube/catheter and wounds, thereby requiring more care and support [19]. Poor functional status is also reported to be associated with aspiration pneumonia [20,21] and urinary tract infection [22]. Thus, before discharge, it is important to ensure that caregivers have good care skills to improve transitional care.   Underlying malignancy instead of CCI score correlates with the use of PDMC and indicates that cancer-associated medications, symptoms, and psychological change like depression may impact on quality of care [23]. However, most PDMCs are without RFS. In contrast, similar to BI, CCI and length of hospital stay significantly correlates with PDMC plus RFS, possibly because these two factors are associated with patient complexity [11,24].
The cost of US$ 5.50 per patient is more expensive than the US report by Rennke et al. of US$ 0.26 [18]. This may be due to the recruitment fee and high call-in rate in Taiwan. The costs may be reduced if this service is integrated into routine transitional care so that no recruitment fee will be required. Although the role of PDMC in transitional care is not investigated in the present study, it can be considered a sensor to detect this problem. A randomized controlled trial or an observational study is needed to investigate the cost-effectiveness of PDMC, especially regarding savings in resources. For those without RFS, telephone counseling may provide patients prompt answers and instructions for problems in a post-discharge care setting. However, for PDMC plus RFS, an effective strategy that can manage RFS should also be studied in the future.
There are three limitations to our study that are worth noting. First, because this study was performed in a tertiary referral center and patients have multiple co-morbidities, whether or not the results can be generalized to regional or district hospitals warrant further investigations. Second, the study uses a telephone service to investigate post-discharge adverse events and this may underestimate the occurrence. Third, the cost of equipment is not included in the present study and may underestimate PDMC cost.
In conclusion, 29% of home-care patients demand for PDMC and 7% have RFS within 30 days. Among the PDMCs, 50% are for help regarding symptom management and the other 50% are for care instructions and administrative services. The use of PDMC is associated with emergency department visits and unplanned readmissions. Pre-discharge care instructions should be emphasized in patients with cancer and low BI. Patients with lower BI, higher CCI, and longer hospital stay are a risk population for RFS after discharge and require other types of support like transitional care.