Sickle cell anemia day hospital: an approach for the management of uncomplicated painful crises

During the period covered in this report (1989-1993), the DH provided care for patients with uncomplicated painful crises. The DH, located in the 1700-square-foot BCSCC at Montefiore Medical Center, included a triage room, 3 beds, and a clinical laboratory for sample processing. The BCSCC also included an outpatient facility for health maintenance; offices for the clinical director, nurse coordinator, administrator, and education and psychosocial counselors; and a meeting/teaching area. The DH hours of operation were Monday through Friday from 9:00 am to 5:00 pm for both walk-in patients and transfer patients from the ED, who were treated in the ED while the DH was closed.

During the first 5 years of the DH, 144 patients sought treatment for acute pain crises of varying frequency. This represents a mean of 70 patients and more than 500 visits per year (Table1). The median age of the patients was 30 years, and 81% of the patients had homozygous sickle cell anemia. Male patients outnumbered female patients during the first 2 years of the facility operation. During the last 3 years of the study, a nearly-equal distribution of the sexes was treated, with female patients averaging 57% of the visits.

We used the following assessment and treatment protocol in the DH. Patients were assessed by a nurse and physician prior to initiation of therapy. Assessment and the initial treatment occurred within 15 to 20 minutes of the patient's arrival at the DH. At half-hour intervals, patients completed the rapidly administered assessment instruments through a sequential combination of nurse interview and self-administered questionnaire. Treatment decisions and final dispositions were made based on the responses. The specific procedures in the DH are listed below.

Treatment is assessment driven. We establish whether the pain is typical or atypical vis-à-vis previous episodes. Pain intensity (overall and at specific sites), pain relief, mood, and sedation are evaluated by employing categorical scales of the Crisis Pain Assessment Form and Visual Analog Scales of the Memorial Pain Assessment Card,30 initially, or the Montefiore modification, a visual numerical scale of the same dimensions, with figure drawings to record the sites of pain. The Brief Pain Inventory Short Form, which also measures the impact of pain on affect and physical functions, is administered before and after the course of treatment.31 32 

Drug selection is based upon the patient's prior history and current assessment. Recording a history establishes the standard drug and dosage required to treat the patient's painful crises, resulting side effects, medications the patient takes at home, and any medication that was taken since the onset of pain. If the patient takes opioids chronically, we assume that the patient is tolerant to the medication, and a different opioid or a higher dose of the same drug is used.

On a scale of 0-4 (0 = none, 1 = little, 2 = moderate, 3 = good, and 4 = complete relief), we define relief as a score of 2 or greater. The effect of therapy is determined in 30-minute intervals after the loading dose. Titration is continued to the highest relief achievable (scores 2-4).

Within 2 hours of titration, maintenance dosing and dosing intervals are determined based upon the dosing required during titration and the duration of relief.

Approximately ¼ to ½ of the maintenance dose is administered for pain that occurs between the scheduled doses. If frequent rescue doses are required, the maintenance dose is adjusted upward or the dosing interval is decreased.

Anti-inflammatory agents, antihistamines, and other adjuvant therapies are used in combination with opioids.17 23 

If, after pain control is achieved, the patient reports that the medication does not relieve pain for as long a period as previously, dose intervals are shortened. If decreased effectiveness is reported, doses are increased or changed to ½ the equianalgesic dose of another opioid, and then titration to relief, maintenance, and rescue are repeated.

Occurrence of nausea, vomiting, pruritus, respiratory depression, and sedation are recorded and treated when indicated. For sedation, the visual numerical scale (0 = wide awake and 10 = asleep) is used, and methylphenidate is employed to reverse or lessen sedation.33 

For example, patient controlled analgesia22,34 35 is indicated if pain is inadequately controlled with bolus titration or if frequent dosing is required for relief maintenance.

Precipitating and propagating events, such as dehydration, acidosis, hypoxia, infection, and stress, also need to be treated.1,3,4-8 Due to hyposthenuria, most patients experiencing painful crises are variably dehydrated. If dehydration is not treated sufficiently with oral hydration, 5% dextrose and ½ normal saline, alternating with 5% dextrose and ¼ normal saline, is given. An alkali is indicated for documented metabolic acidosis, and oxygen is given for documented hypoxia over steady state.36-38 A low mood score after pain is relieved or bizarre graphic rendition of pain indicates a high possibility of psychological stress.30 39-41 The patient is questioned further, basic stress reduction techniques are employed, and referrals are made for follow-up visits.

After patients are treated in the DH, they are discharged home, transferred to the ED, or hospitalized. If pain is controlled and can be managed at home, small prescriptions of a graded analgesic regimen, based on severity of the acute event, are prescribed for patients at discharge, and an appointment for prompt follow-up is made.3,17,18 41 Patients are referred to their primary health care providers for health maintenance and routine home care. If pain is not significantly diminished, if parenteral therapy is required to maintain relief, or if significant comorbidities are present, then the patient is hospitalized. If an inpatient bed is not available or if the DH is closing for the day, the patient is transferred to the ED to continue treatment.

Using the DH database and the hospital's information systems, we compared the DH visits and dispositions of painful crises with those at the ED. The ED admission rate in the year prior to the opening of the DH was taken as baseline. The ED also treats the majority of patients presenting with comorbidities as well as patients with uncomplicated pain. To adjust for this, we compared DH visits only with the portion of the ED population that is comparable to the uncomplicated painful crisis population which constitutes the majority of the patients seen by the DH. Patients admitted from each unit without comorbidities, at admission and during the hospital course, were identified from inpatient discharge data. Those patients who were admitted with only unrelieved pain were combined with patients who were discharged home from the ED or the DH in order to identify the population of patients with uncomplicated pain for each unit. Admission rates for uncomplicated pain were then compared.

From the hospital admission and discharge database, the LOS for adult painful crisis patients without comorbidities (who were followed by DH physicians with house staff assistance) was compared with the LOS for patients followed by physicians unassociated with the DH.

Descriptive statistics were employed to evaluate pain intensity, pain relief, and hospitalization rate (SAS statistical package). Student's t test was used to determine significance of relief and time to relief of high (frequent visits) and low (infrequent visits) users of the DH.

We also estimated the costs of operating the DH and the cost savings for reducing hospital admission rate and LOS for patients without comorbidities.42 We used $1200 as the average daily cost of a medical/surgical bed, and an average length of hospital stay was estimated at 9.3 days. A cost-reduction model was used to calculate the impact of the decreased rate of hospitalizations. The savings of decreased hospitalization costs were calculated based on a diagnosis-related group (DRG) of 4.3 days. We estimated conservatively that 7 days of hospitalization costs were covered by third-party payors and that the 1.5-day excess was a loss.

Inpatients followed by DH staff had shorter LOS hospitalizations, and these savings were also calculated by multiplying the reduction in LOS by the number of admissions and the daily cost of occupying a hospital bed.

There were 2554 visits to the DH (Table 1) over the 5-year period. Of these visits, 2257 were walk-in patients, and 297 patients were transferred from the ED (Figure 1). During years 1 and 2, 76% of visits were walk-in patients. Walk-in visits increased to 94% for years 3-5. As shown in Table2, 22% of the patients accounted for 67% of total DH visits.

Patients reported moderate pain in 40% of the DH visits and severe pain in 60% of the visits. On a scale of 0-3, as described previously, the overall mean pain score was 2.7 (SD = 0.7), and the median pain score was 3.0. Detailed results of pain assessment in the DH will be presented elsewhere. Patients were treated within 20 minutes of arrival. Since opioids were selected based on prior analgesic history, initially drug usage included meperidine (Demerol, 90%), morphine (4%), hydromorphone (Dilaudid, 3%) and levorphanol (Levo-Dromoran, 3%). Opioid usage was modified over time (Figure2), and by the fifth year, usage included meperidine (63%), hydromorphone (33%), morphine (2%), and levorphanol (2%). Combination therapy was used during 75% of the visits. Opioids were used in combination with antihistamines (72%), nonsteroidal anti-inflammatory drugs (NSAID, 17%), and both antihistamines and NSAID (11%). In the third year of the study, ketorolac was introduced to the treatment regimen in 20% of the visits, and it was given to patients who historically were poor responders or who achieved less than good pain relief during the course of therapy.

Over the 5-year period, 84% of patients were titrated with medication to pain relief, and 40% of these patients had relief within 1 hour (Figure 3). Those patients with unrelieved pain (16%) were frequent-pain patients who experienced more than 5 visits and more than 2 hospitalizations per year. The overall mean relief score on a scale of 0-4 was 2.5. For frequent-pain patients, the mean relief score was 2.20 (SD = 0.4), whereas for infrequent-pain patients, the mean relief score was 3.1 (SD = 0.7); this difference was statistically significant (P < .0001). The overall mean time to relief was 2.5 hours. For frequent-pain patients, the mean time to relief was 3.4 hours (SD = 1.2), whereas for infrequent-pain patients, the mean time to relief was 1.7 hours (SD = 0.7); this difference was statistically significant (P < .0001). Some frequent-pain patients consistently reported pain relief only at the end of the visit (during the last assessment).

Side effects, such as nausea, pruritus, vomiting, or sedation, were experienced in 20% of the patients treated in the DH. Pruritus occurred most often following administration of hydromorphone. Seizures, clonus, or respiratory depression were not observed.

On average, 80.5% of DH visits resulted in the patient being discharged home; 8% of the visits were admitted to the hospital, and 11.5% were transferred to the ED. The percentage of walk-in DH patients discharged home increased from 70% in the first 2 years to 90%-94% in the last 3 years (Figure 1).

During the 5-year study, there were 2259 walk-in visits to the DH and 2401 direct walk-in visits to the ED (Figure4). In the ED, 25% of patients seeking care accounted for 70% of ED visits (more than 6 visits per year). During these 5 years, 91% of the ED visits were by patients who also had visits to the DH.

The average LOS in the DH was 4.5 hours (range, 2 to 7 hours), while the average LOS in the ED was 13 hours (range, 11 minutes to 90 hours). During the 5-year period, the average LOS in the ED remained fairly constant for the dispositions of admissions (13-16 hours) and discharge home (11-13 hours). Treatment time in the ED before transfer to the DH decreased each year, ranging from 16 hours in year 1 to 8 hours in year 5.

During the 5-year study, an average of 51% of the ED visits (70% in the first year) and 8% of the DH visits were admitted to the hospital. A baseline assessment of ED dispositions during the year prior to establishment of the DH revealed an admission rate of 92% (Figure 4). Overall, there was a 40% reduction in hospital admission rate compared to the baseline rate. Dispositions according to frequency of visits show that the DH was more successful in discharging patients home, particularly in the high-user category (Figure5).

To correct for potential differences in severity between patients seen in the ED and DH, patients without comorbidities were compared. For each facility, patients discharged home and patients admitted without comorbidities were combined to represent their respective uncomplicated pain populations. The majority of visits, 70% (1818) of the ED visits and 80% (2033) of the DH visits, were for uncomplicated pain. Of the patients with uncomplicated pain, 776 out of 1818 ED patients (42.7%) were admitted to the hospital, and 168 out of 2033 DH patients (8.3%) were admitted.

The possibility of recidivism (DH patients discharged home who promptly sought medical care in the ED or our facility) was examined in the first and fifth years of the study. Within 3 days of discharge from the DH, 9.5% of the discharges resulted in a revisit to either facility. All patients involved were frequent-pain patients (17.8% of patients accounted for 82% of these revisits). Of these revisits, 21% were admitted to the hospital.

Whether patients were admitted through the DH or ED, LOS for adult patients with painful crises without comorbidities (and who were followed by DH physicians, with the assistance of the house staff) was reduced from 9.3 days in the first year to an average of 7.3 days in the fifth year (average 7.8 days per year over the 5-year period). The LOS for patients followed by private physicians who were not connected with the DH remained unchanged.

In the 5 years reported, the DH spent $468,917 and billed $403,861, which would lead to an assessment of the DH, at least initially, as a money-losing operation. Nevertheless, the presence of the DH generated savings for the institution by reducing the hospitalization rate of patients with outlier days and by actually decreasing the LOS of inpatients, as described below.

The amount of savings due to reduction in hospitalizations was calculated using a DRG of 4.3. During the first year of the study, the baseline LOS was 9.3 days, and for inpatients followed by DH personnel, this figure was reduced on average to 7.8 days for the total 5 years. We conservatively estimated that with these patients, 7 days of hospitalization left the hospital without a deficit. Calculating the 1.5 day outlier balance as a loss per admission for 700 saved admissions, and using $1200 as the average daily hospitalization cost for a medical/surgical bed, we accrued an estimated savings of $1,200,000 compared with a situation in which all patients went to the ED.

Inpatient admissions followed by DH personnel had a shorter admission stay than those same patients followed by non-DH personnel. This resulted in further savings of $450,000.00 (250 admissions × 1.5 days × $1200/d).

The grand total savings, therefore, amounts to more than $1.7 million, which makes this operation financially sound. Of note is the fact that we cannot assess in monetary terms the impact of a less crowded ED and the decrease in suffering due to more rapid control of pain.

Our experience in the DH demonstrates that most painful crises can be managed successfully and that hopelessness in the pain treatment of these patients, including those with frequent pain, is unwarranted. After a slow start, pain was ultimately controlled in 90% of the patients, a rate similar to other pain states and a more rapid control than generally achieved in sickle cell patients.35 

The underpinning to this approach is immediate assessment and treatment. Each subsequent treatment decision was preceded by and based upon assessments of pain, psychological distress (mood), response to therapy (relief), and adverse effects (sedation). Medication titration to pain relief individualizes care through prompt reassessment and treatment at regular frequent intervals. This approach is preferable to (1) customary rote prescription of the same drug, dose, and interval for each event; (2) as-needed dosing, which permits pain to recur before treatment is provided; or (3) prejudicial exclusive use or avoidance of a particular analgesic.17,23 43 

There is bias against opioid use in pain management, and this bias is a well-recognized factor in undertreatment, even before drug preferences present obstacles to management. These biases are amplified in certain racial groups44,45 and in persons who have frequently recurring acute and/or chronic pain.46-48 Data show that African Americans and Hispanic Americans are undermedicated for their pain to a greater extent than their Caucasian counterparts.44 45 Our personnel were particularly sensitized to these issues because the majority of our patients with sickle cell anemia are either African American or Hispanic American (Table 1), the mainstay of therapy is opioid analgesics, and a few patients account for most of the pain visits.

The choice of a drug was based on prior analgesic history and current assessment. Opioids are not inherently more or less efficacious when used to treat a group of patients, but they are highly influenced by pharmacokinetic and pharmacogenetic determinants of opioid response in individuals.49 Hence, treatment has to be individualized. Furthermore, any ineffective opioid administered chronically to an individual is potentially injurious.50 ⁵¹ Therapy modification included a reduction in meperidine usage (adjustment for tolerance), although none of the patients had seizures from meperidine usage in the DH. Adjuvant therapies, including antihistamines, NSAID, and introduction of parenteral ketorolac in year 3, were found useful.

Undertreatment of pain in medicine is well recognized,16,21,52 in part as a consequence of bias to opioid usage, especially in minority and frequent-pain patients. Furthermore, it is possible that if pain control is delayed, the undertreatment becomes a propagating factor. The sequelae of untreated or undertreated pain include hypoxia, dehydration, acidosis, and stress,29 52 all well-known precipitating factors for painful crises. By addressing the great variability of patients' responses to therapy and promptly assessing and titrating the patients' treatments, we decreased the magnitude, duration, and time to relief of pain. More than 80% of patients were titrated to relief during the DH visit, and this time period was less than 4 hours for frequent-pain patients and less than 2 hours for nonfrequent-pain patients. The increase in relief and pain reduction in the DH is at least partially attributed to an increase in walk-in visits (from 80% to 99%) and a corresponding decrease in transfers from the ED, resulting in greater initiation of therapy in the DH. Reduction in meperidine usage and the addition of parenteral ketorolac coincided with a decrease in the rate of hospitalizations.

Since the DH opened, there has been a 40% reduction in hospital admission rate compared with the dispositions in the ED when it was the only site for treatment. Overall, there was a 5- to 7-fold decrease in the rate of hospital admissions by the DH compared with admissions by the ED. A decreased hospitalization rate was observed among patients with uncomplicated painful crises who were transferred from the ED to the DH, although this decrease was not of the magnitude of the decrease for direct walk-in visits. Although the ED saw more patients with comorbidities (30% ED compared with 20% DH), when the rate of admissions of painful crises without comorbidities were compared, the rate of admission remained 5-fold less in the DH.

While we were able to reduce the hospitalization rate of frequent-pain patients, they remain a problem and require a supplemental approach.53 In the DH, 22% of patients accounted for 67% of the visits, and this proportion was similar in the ED.

The average length of hospital stay decreased during the 5-year study among adult patients followed by the DH staff, but the length of hospital stay remained stationary in patients followed by staff not associated with the DH.

The DH-associated savings were generated by decreasing the time in the outpatient facility, the rate of hospitalization, and the inpatient length of hospital stay.54,55 The hospital administration response to the DH performance has been encouraging. After years of uninterrupted federal funding of this demonstration project, an economical analysis resulted in the complete funding of an expanded DH operation by the hospital. In our new facilities, we now have 6 beds (rather than 3), expanded personnel, and expanded hours of operation (from one to two 8-hour shifts). This is the best proof of the economical soundness of this operation. Other institutions have also supported successful dedicated facilities for sickle cell anemia patients.56 57 

A major dilemma and current focus in pain management is to find ways to actually change practices.58-60 Despite the climate for change created by research advances, evidence-based clinical practice guidelines, quality improvement, and educational activities, traditional patterns of professional practices have persisted as the most formidable barrier to relieving pain.61 Our results buttress the concept of alternative care delivery systems, such as the dedicated DH, as a pivotal element for removing such barriers. Not only is it a centralized unit that treats pain effectively, it also represents a resource for patients, families, and health care professionals; serves to rapidly incorporate research advances; and has a clear impact on in-house care because patients who are hospitalized are already in an appropriate track of treatment.

In conclusion, a dedicated facility for the treatment of uncomplicated painful crises, operating on principle-based pain management, can reduce the time to pain relief, increase the number of patients discharged home, decrease the hospitalization rate, lessen the use of the ED, help integrate care, and positively impact use in other areas. Hence, a dedicated DH appears to be an advantageous multidisciplinary alternative care facility for the acute care of uncomplicated painful crises in sickle cell disease patients. The approach presented is applicable in various clinical settings. The most critical elements include having a dedicated facility and leadership adapted to the various types, sizes, or locations of the institution (urban, suburban, or rural).

The authors wish to acknowledge the nurses, hematology fellows, and staff of the Sickle Cell Center; the medical house staff; and members of the ED, all of whom contributed significantly to the success of this effort. Institutional support is gratefully acknowledged.