General Medical Care External Hospitalizations for Patients in Texas State Mental Health Hospitals
The Journal of Texas Medicine – March 2013
Tex Med. 2013;109(3):e1.
By Alan B Shafer, PhD; Ryan Kumar Ray, MPH; and Emilie A. Becker, MD
Dr Shafer, Mental Health and Substance Abuse Division, Texas Department of State Health Services, Austin, Texas; Mr Ray, The University of Texas School of Public Health, Houston, Texas; and Dr Becker, Mental Health and Substance Abuse Division, Texas Department of State Health Services, Austin, Texas. Send correspondence to Alan Shafer, PhD, Mental Health and Substance Abuse Services Division – Decision Support, Texas Department of State Health Services, 909 W 45th St, Building 634, Mail Code 2144, Austin, TX 78714-9347; email firstname.lastname@example.org or email@example.com.
We explored the characteristics of general medical hospital admissions for patients in state mental health hospitals. Data were extracted from a statewide database of all hospital discharges for 5 years identified as general medical hospital admissions that occurred during the stay of patients at state mental health hospitals. Across the 9 mental health hospitals in the state system, rates of admission to general medical hospitals varied significantly from 0.7% to 3.7%. On average, of the 1.9% of all state mental health inpatients who had a general hospital admission, 25% occurred within 4 days of admission to the mental hospital. The average general hospitalization lasted 5.7 days. The reported total charge for all stays was $34 million. Dehydration (15%), hypertension (10%), and diabetes (10%) were the most frequent diagnoses. Thirteen percent of diagnoses met preventable hospitalization criteria. Given the variability among hospitals in admission rates and the number of preventable conditions, improvements in patient care and health as well as reduced admissions to general medical hospitals may be possible.
Rates of psychiatric comorbidity, especially depression, among the medically ill have been examined frequently and have generated several meta-analyses. However, the reverse question of how often medical illnesses occur in the mentally ill has been less frequently examined. In particular, the rate of medical illness among hospitalized psychiatric inpatients has been rarely explored.
A number of studies have examined outpatient rates of medical comorbidity among psychiatric patients. One recent literature review of general practice1 indicates that schizophrenic patients have a relatively high level of general medical comorbidity. The most common diseases of these patients were metabolic syndrome (36%), chronic obstructive pulmonary disease (COPD) or asthma (22%), and diabetes (9% to 14%). Findings show that the greater the number of current medical problems, the greater the severity of psychosis or depression.2 Bipolar patients also suffer from high levels of medical comorbidity.3 Rates of comorbidity increase with age, but overall, approximately 30% to 50% of bipolar patients younger than 70 years have one or more major medical comorbidities. The most frequent diagnoses were cardiovascular-hypertension (10.7%), COPD (6.1%), and diabetes (4.3%). Along with schizophrenic and bipolar patients, depressed patients also suffer from high rates of comorbid medical illness, with estimates ranging from 10% to 40%.4
Relatively few studies have examined medical comorbidity among psychiatric inpatients. One study found that approximately 12% to 15% of inpatients had a serious medically active condition that was a focus of care during their stay.5 These patients on average stayed three days longer than comparable patients without medical comorbidity. Another recent study6 found that approximately 2% of psychiatric inpatient admissions were transferred to a medical-surgical unit within 48 hours of admission. Most of these transfers were unforeseen and unpreventable, where medical symptoms and conditions did not occur until after admission into the psychiatric unit.
The extent of illness and associated costs among patients in Texas state mental health hospitals has not previously been systematically explored. This study explored the number and cost of external hospitalizations in which these patients were sent to general medical hospitals for care. The primary goals were to determine the number of patients, the costs, the most frequent conditions, and if these hospitalizations were preventable. The secondary goal, based on these results, was to suggest what could be done to improve future care of patients.
The Texas Health Care Information Council's Texas Hospital Inpatient Discharge Research Data Files for 2005-20097 were the sole data source. The Hospital Inpatient Discharge File consists of hospital discharge data from all Texas licensed hospitals, except the few hospitals statutorily exempt from the reporting requirement. Based on internal Department of State Health Services (DSHS) records, nine DSHS mental health hospitals sent their medically ill patients to 34 general hospitals, whose records formed the initial data set.
Only adult (18 years and older) patients’ records were retained for the study. Records were matched on the basis of two criteria. First, the unique patient identification number in both the state mental health hospital and the general hospital had to match. Second, the date of admission into the general hospital had to fall within the dates of admission and discharge from the state mental health hospital. The date of discharge from the general hospital could occur after the date of discharge from the state mental health hospital.
A total of 1268 matching discharge records identified for 1021 unduplicated unique patients represented 510 men and 511 women; 430 whites (45%), 375 Hispanics (30%), 152 blacks (15%), and 64 patients (6%) of some other race or ethnicity. The average age of these patients was 41.6 years (SD = 14.7). Approximately 27% of the discharges were for patients younger than 30 years, 27% for patients aged 50 to 64 years, and 7% for patients older than 65 years.
From 2005 to 2009, state mental health hospitals admitted 52,799 unique patients with a total of 81,983 discharges. Of those state mental health hospital patients, 1.9% or 1021 unique patients had a total of 1268 discharges from one or more of the 34 general medical hospitals that the state used to provide general medical care to psychiatric inpatients. Most patients had either 1 (82%) or 2 (13%) discharges from general hospitals, fewer patients had 3 discharges (3.8%), and very few patients (< 1%) had 4 or more discharges from general hospitals.
The patients sent from state mental health hospitals to general medical hospitals were slightly older (M = 41.5 years) than most state mental health hospital patients (M = 35 years) (t(1059) = 13.99 and P<.01).
Females in state mental health hospitals (42%) were more likely to be sent to a general medical hospital (50%), while males (58% of the overall state mental health hospital population) were less likely to be sent to a general medical hospital (50%) (chi square  = 24.4, P<.01).
Hispanics (36% general medical hospital admissions versus 27% state mental health hospital population) and patients of other races and ethnicities (6.3% general medical hospital admissions versus 2.2% state mental health hospital population) were more likely to be sent to a general hospital; while blacks (15% general medical hospital admissions versus 19% state mental health hospital population) and whites (42% general medical hospital admissions versus 50% state mental health hospital population) were slightly less likely to be sent to a general medical hospital (chi square  = 129.7, P<.01).
Timing of Admissions
The average timing of admission to a general hospital from the state mental health hospital was 37 days after admission to a state mental health hospital (SD = 54). However, this distribution is skewed with 50% (median) being sent to a general hospital within 15 days of admission and 25% being sent within 4 days of admission. Additionally, a relatively large number of patients (8.75%) were sent to a general medical hospital on the day of admission to the state mental health hospital, and an additional 6.7% were transferred on the day after admission to the state mental health hospital.
Source of Admissions
A significant difference was found among the 9 state hospitals in terms of the number of observed admissions to general medical hospitals (chi square  = 321, P<.01). The average percentage of unique patients sent to general medical hospitals was 1.8% but ranged from a low of 0.7% at a small rural state hospital to a high of 3.7% at a larger state hospital serving both an urban area and forensic patients. Analyses of individual hospitals revealed that the 3 smallest hospitals’ ratios of observed to expected admissions to general medical hospitals was not statistically significant. In other words, they send approximately as many patients to hospitals as would be expected based on their size. For example, if a state mental health hospital served 10% of the states' mental health patients, we would expect it to also have about 10% of the patients sent to general medical hospitals. The overall differences were driven by the 6 largest hospitals, which account for 85% of the general medical hospitalizations. Among these 6 largest hospitals, 2 had fewer than expected general medical hospitalizations. The remaining 4 had greater than expected general medical hospitalizations, which accounted for 69% of the total general medical hospitalizations.
Length of Stay and Cost
The average length of stay in the general medical hospital was 5.78 days (SD = 7.36), with 50% of patients staying 4 or fewer days and 90% staying fewer than 10 days. The average charge for each of these general medical hospitalizations was $27,006 (SD = $56,858). The total charge for all patients for all stays in general medical hospitals was $34,243,358.
The average charge per patient (not stay) across hospitals ranged from $17,900 to $46,400. The total charges for general medical hospitalizations among state hospitals ranged from below $500,000 for a small rural hospital to more than $10,000,000 for a large hospital serving both urban and forensic patients. Of the 34 general medical hospitals that admitted these patients, the top 5 accounted for 60% of the charges, and the top 10 accounted for 85% of the charges.
The Hospital Inpatient Discharge Files used for this study contained up to 4 diagnoses. Of the 1268 individual general medical hospital discharges, a total of 3718 had nonmental health diagnoses (all discharges had multiple diagnoses). Examination of the frequency of all diagnoses (across all 4 diagnostic fields) showed that the top 25 three-digit ICD-9 diagnostic group codes accounted for 1861 (50%) of all the diagnoses. The top diagnostic group was ICD-9 276: Disorders of fluid, electrolyte, and acid-base balance, with 201 (15%) of general medical hospital discharges having this diagnosis. Outcome of delivery (V27) was the second most common diagnostic group with 145 patients. The next 2 most common conditions were ICD-9 140: Essential hypertension and ICD-9 250: Diabetes mellitus. About 10% of the general hospital discharges had these latter 2 diagnoses. The Table displays the total diagnoses ordered by cost. Other common medical conditions appeared also in the top diagnostic groups, including circulatory (heart failure, cardiac dysrhythmias, and other forms of chronic ischemic heart disease) and respiratory conditions (pneumonia, asthma, chronic airway obstruction, not elsewhere classified), as well as abscess, urinary tract infections, renal failure, and anemia.
The largest overall cluster of diagnostic groups (among patients having multiple diagnoses) related to pregnancy and delivery. A total of 6 diagnostic groups in the top 25 related to pregnancy and delivery accounted for about 11% of the total hospitalizations. However, the charges for deliveries were relatively modest, and all of them were at or near the bottom of the top 25 diagnostic groups in terms of cost.
Nine potentially preventable hospitalization conditions as defined by the Agency for Healthcare Research and Quality8,9 were examined also. Preventable hospitalizations refer to patients who are admitted to a hospital for treatment with a diagnoses or condition that could have been avoided had they received proper outpatient care earlier. Note that preventable hospitalizations are defined by the primary diagnosis, not any diagnoses (such as secondary and tertiary) as described in the previous section. Furthermore, only a specific subset of designated primary diagnostic codes define a potentially preventable hospitalization.
Of the total 1268 admissions, 148 (13%) had primary diagnoses classified as potentially preventable. The total costs for these admissions were $3,593,995, about 12% of the total costs. The top 5 potentially preventable hospitalizations were for bacterial pneumonia (43 admissions), diabetes (28 admissions), dehydration (22 admissions), urinary tract infections (16 admissions), and congestive heart failure (15 admissions). These accounted for most of the potentially preventable hospitalizations (124 admissions total, about 84%) and most of the costs ($2.8M, 87%) of these potentially preventable hospitalizations. The remaining potentially preventable hospitalizations, including hypertension, chronic obstructive pulmonary disease, angina, and asthma, had fewer than five admissions.
Overall, the rate of admission to general medical hospitals from state mental health hospitals was fairly low, with < 2% of all patients requiring external care. However, given the variability across hospitals and that approximately 13% of general hospital admissions were potentially preventable, opportunities exist to improve patient health and reduce costs associated with these hospitalizations. Even a small reduction in admissions could result in cost-savings. For example, averting just half of the preventable hospitalizations could save approximately $350,000 annually.
Similar to previous studies on outpatients, some of the most common diseases and conditions among state mental health hospital patients referred for outside medical care were chronic obstructive pulmonary disease, diabetes, cardiovascular conditions, and hypertension. Most previous studies focused on chronic conditions, but acute care conditions were also some of the top diagnostic groups in this study, including dehydration, acute exacerbation of chronic renal failure, pneumonia, and urinary tract infections. Many of these admissions, while acute, are also for potentially preventable conditions that may be the most amenable to health care intervention.
Demographic differences existed between the patients sent to general medical hospitals compared with the other patients in the state mental health hospitals. Medical diagnoses increase in number with patient age, so finding that patients sent to general medical hospitals were generally older was expected. In terms of gender differences, if deliveries are excluded, then the percentage of males and females sent to general medical hospitals are almost exactly the same as the state mental health hospital population. This suggests that the deliveries explain the higher rates of admissions for females. The reasons for race and ethnic differences are unknown. Hispanics were sent to general medical hospitals more and blacks less than in the overall state mental health patient population. Further research on this particular difference is needed. The reasons why “other” race and ethnicity patients were sent for care more frequently are unknown. This category includes a number of specific subgroups and had too few patients (only 64) to examine to any great degree.
The major limitations of this study concern the data, which are collected primarily for administrative rather than research purposes. Although the impact is likely to be minimal, the anonymous unique identification provided with the data set has accuracy limitations. The hospital charges recorded in the data file do not reflect the actual sums paid; rather, they represent the initial charges submitted by the hospital to the payer. Therefore, the final actual amount paid for services reported in the results can be inferred as less, but how much less is unknown. Patient confidentiality rules prevent using the hospital discharge data set to identify specific patients to corroborate results and examine actual medical records.
Despite these limitations, this study has provided useful information to hospital management and the trends between the rates of referrals to general medical hospitals are generally accurate.
Some suggestions to improve care are as follows:
- Develop standardized protocols for when patients should be sent to a general medical hospital. At this time, no referral protocols exist for any medical conditions at any of the 9 state mental health hospitals.
- Provide additional on-site or telemedical care with extended hours in state mental health hospitals to reduce night and weekend admissions.
- Implement proactive screening of complex medical conditions before state hospital admission to reduce the numbers of inappropriate admissions of patients too ill to be in a psychiatric facility.
- Establish contracts with local urgent care providers to allow treatment of those patients too sick for a psychiatric hospital but not sick enough for a tertiary care hospital.
- Have more uniform data collection by each state hospital to allow for more timely comparisons between them. Included in this collection would be disease outcome measures to make sure those who need outside medical care indeed receive it.
- Educate state hospital physicians about diagnosis and treatment of preventable causes of hospitalization to target that sector of medical admissions.
In summary, medical costs of state hospital inpatients remain understudied and further analyses could likely provide significant savings as well as better, more timely, and higher-quality health care.
- Oud MJ, Meyboom-de Jong B. Somatic diseases in patients with schizophrenia in general practice: their prevalence and health care. BMC Fam Pract. 2009;10:32.
- Dixon L, Postrado L, Delahanty J, Fischer PJ, Lehman A. The association of medical comorbidity in schizophrenia with poor physical and mental health. J Nerv Ment Dis. 1999;187(8):496-502.
- Beyer J, Kuchibhatla M, Gersing K, Krishnan KR. Medical comorbidity in a bipolar outpatient clinical population. Neuropsychopharmacology. 2005;30(2):401-404.
- Goodnick PJ, Hernandez M. Treatment of depression in comorbid medical illness. Expert Opin Pharmacother. 2000;1(7):1367-1384.
- Lyketsos CG, Dunn G, Kaminsky MJ, Breakey WR. Medical comorbidity in psychiatric inpatients: relation to clinical outcomes and hospital length of stay. Psychosomatics. 2002;43(1):24-30.
- Passov V, Rundell JR. Analysis of transfers from a medical-psychiatry inpatient unit to a medical-surgical unit within 48 hours of admission. Psychosomatics. 2008;49(6):535-537.
- Texas Department of State Health Services. Texas Hospital Inpatient Discharge Research Data File: 2005-2009. Austin, TX: Texas Department of State Health Services, Center for Health Statistics; 2012.
- Agency for Healthcare Research and Quality. Refinement of the HCUP quality indicators. Technical Review No. 4. AHRQ publication No. 01-0035. Rockville, MD: Agency for Healthcare Research and Quality; 2001.
- Kruzikas DT, Jiang HJ, Remus D, Barrett ML, Coffey RM, Andrews R. Preventable Hospitalizations: A Window Into Primary and Preventive Care. HCUP Fact Book No. 5. AHRQ Publication No. 04-0056. Rockville, MD: Agency for Healthcare Research and Quality; 2004.
March 2013 Texas Medicine Contents
Texas Medicine Main Page