Early discharge predictors among inpatient crack cocaine users

Abstract Introduction High rates of early hospital discharge are often observed in crack cocaine users and are related to adverse outcomes and increased public spending. This study evaluated clinical and sociodemographic factors associated with early treatment discharge among crack users. Methods The sample comprised 308 men diagnosed with crack cocaine use disorder (crack only), aged 18 to 65 years, admitted between 2013 and 2017 to a male-only hospital unit to treat substance use disorders. Sociodemographic and clinical data were obtained using the Addiction Severity Index, 6th version, and a Sociodemographic Questionnaire. Results Early discharge (within 7 days) was significantly associated with lack of own income, insufficient family support, being single, and recent homelessness. Regarding drug use, lower treatment retention was related to younger age of crack use onset, recent alcohol use, and nicotine use. Factors such as age, skin color, and educational level showed no relation to the outcome. Conclusion Our findings suggest that presence of characteristics verifiable at the time of admission may be related to crack users’ treatment retention. Identification of these factors can contribute to target interventions in order to improve treatment adherence in crack cocaine users.


Introduction
Epidemiological studies have shown increased crack consumption (cocaine smoking) in Brazil over recent years. 1,2 Research estimates that nearly 0.7% of the adult population and 35% of illicit drug users living in state capitals regularly smoke crack. 1 Crack dependence is related to higher mortality rates due to violence, to higher incarceration rates, and to behaviors that increase the occurrence of negative health outcomes, such as risk behaviors for human immunodeficiency virus (HIV) and viral hepatitis contamination 3-7 and attempted suicide. 8 In a 12-year follow-up study with crack users, Dias et al. observed high mortality (20%) and incarceration (10%) rates. 9 Studies have also observed violent behaviors and involvement in crime in these users. 7,[9][10][11][12] Although the prevalence of crack use is lower than that of other drugs, such as tobacco, alcohol, and cannabis, this substance is responsible for most psychiatric hospitalizations in Brazil. 6,10,12,13 Detoxification hospitalization is an important element of treatment; besides helping manage initial withdrawal symptoms, the stabilization stage enables motivational strategies to be developed, which can improve therapeutic adherence. 14,15 Nevertheless, individuals with crack dependence have high rates of early discharge requests in voluntary hospitalizations in detoxification units. [16][17][18] Previous studies indicate that the rate of early withdrawal after psychiatric hospitalization ranges from 3 to 82%, with the highest percentages found among drug addicts. 16,19 This abandonment is related to higher readmission rates, 20,21 and higher mortality rates from suicide and violence. 16,17,22 A recent Canadian study observed that the readmission rate among illicit drug users discharged early (within 7 days of hospitalization or less) is 50% in one year. 23 Studies conducted in North America and Europe identified several factors associated with low retention during detoxification hospitalization in drug users, such as number of previous admissions, unfavorable employment status, 20 involvement in crime, 20,24,25 being intoxicated at the time of admission, being a racial minority in the hospitalization unit, being younger, being male, 24 and early drug use onset. 26 Factors related to greater social vulnerability and greater severity of dependence such as irregular housing or homelessness, disrupted social ties, and lack of access to social care services were also associated with this outcome. 18,27 Although previous studies have investigated factors related to treatment retention, no Brazilian studies have addressed this issue in the social, cultural, and clinical context of crack users. Unveiling the predictive factors of early discharge can help to identify individuals more susceptible to this outcome and those who require personalized interventions to develop more effective treatment strategies. Pertinent information that can be easily assessed at the time of admission may be an essential factor concerning treatment retention.
As such, this study aimed to assess which social, demographic, and clinical factors, verifiable at the time of admission, may be related to low treatment retention in male crack users.

Design and sample
This cross-sectional study was approved by the

Statistical analysis
The main outcome was the length of hospital stay, defined by the time, in days, that subjects remained in treatment at the hospital. Besides medical hospital discharge, after stabilization of the patient's clinical condition, discharge could be requested by patients themselves or could be recommended by the hospital staff (usually for administrative reasons). Length of hospital stay was divided into quartiles (minimum 1 day -maximum 76 days). The first quartile was equivalent to 7 days of hospitalization or less and considered the early discharge group. To identify predictors of early discharge, we compared this group with the group of subjects that remained 8 days or more hospitalized in the detoxification unit. We selected possible predictors taking into account findings from previous studies conducted in other countries that had evaluated factors related to early discharge of drug users. 18,20,[24][25][26][27] Quantitative variables with normal or symmetric distribution were expressed as mean and standard deviation and compared between groups using the t test for independent samples. Asymmetric quantitative variables were expressed as median and interquartile interval and compared using the Mann-Whitney test.
Categorical variables were expressed as absolute and relative frequencies and underwent the chi-square test of association. All analyses were performed on IBM SPSS® version 18 software, using a 95% confidence level.

Results
Mean hospital stay was 18.7±14.8 days. Early discharge, which occurred within 7 days of admission or less, accounted for 27% of the hospitalizations evaluated (n = 83). Table 1  The results show that factors related to the presence or quality of interpersonal relationships were also associated with the outcome (Table 2), including spending time with family members or close friends (p = 0.044) and talking to people close to them about their feelings and problems (p = 0.049). Hospitalizations that occurred in the colder months (June, July, August, and September) showed a tendency to longer duration (p = 0.060). Difficulty in controlling temperament also seems to be related to shorter hospital stays (p = 0.060). Variables related to legal status and/or criminal activities, religious activities, and having HIV showed no significant differences between groups.    Information on the history and profile of substance use also provides important data regarding retention on a detoxification treatment, indicating a need to adapt the initial therapeutic plan for specific cases. 16,33 In this context, greater attention seems to be needed

Conclusions
Our findings suggest the existence of characteristics verifiable at the time of admission that may be related to crack users' treatment retention. Early identification of these factors may allow for optimization of therapeutic strategies. Decisions as to which setting, outpatient or hospital, is more appropriate for detoxification treatment can be based on these characteristics.
Regarding admission to specialized hospitals, our results identify intervention priorities to improve retention and decrease early discharge rates.