Department, the BC Coroner's Office, the Vancouver Detox Center, the Vancouver City Police Jail for ... centers, and 911 calls with welfare check distribution in.
118
ACADEMIC EMERGENCY MEDICINE
FEB 1997 VOL 4/NO 2
Mortality and Morbidity Associated with the Distribution of Monthly Welfare Payments Glenn Verheul, MD, Sharon Manson Singec PhD, James M. Christenson, MD
I
ABSTRACT .................... ........................
........................
......................
. . .. . . .
Objective: The impact of major social policy decisions on community health is rarely considered or analyzed. This article describes the association of major community and health resource use in relation to the distribution of monthly welfare payments. Methods: A descriptive, retrospective study was performed using existing accessible databases in the city of Vancouver, British Columbia (BC), and St. Paul’s Hospital, a tertiary care, downtown institution. The mean numbers of admissions or responses per week and per day related to the monthly welfare check issue day in 1993 were collected from the following health agencies: the BC Ambulance Service, the Vancouver Fire Department, the BC Coroner’s Office, the Vancouver Detox Center, the Vancouver City Police Jail for public drunkenness, and St. Paul’s Hospital ED. Results: Comparison of weekly events for non-payweeks vs the week starting on welfare payday (mean I+SD) are: St. Paul’s ED, 949 5 51 vs 993 -t 81 (p = 0.10); Detox Center observation admissions, 29 f 5.6 vs 40 +: 7.3 (p < 0.001); Vancouver Fire Department medical responses, 453 2 44 vs 527 ? 45 (p c 0.001); BC Ambulance Service responses, 3,338 2 101 vs 3,634 2 85 (p < 0.001); and coroner-reported deaths, 8.8 2 3.0 vs 13.6 -I 2.6 (p < 0.0001). Conclusions: As measured in multiple independent databases, there is a significant increase in morbidity and mortality in the week after the distribution of monthly welfare paychecks. Key words: alcohol; intoxication; injury; payday; welfare payments; social policy; health policy; public policy; welfare payments; emergency medical services. Acad. Emerg. Med. 1997; 4:118-123.
The numbers of health and social policy decisions set in isolation may be reduced through a growing recognition of the social determinants of population health,’-5 specifically that higher annual income is positively correlated with good health. The 1993 disbursement for social assistance in British Columbia (BC) was $1.67 billion. In . .......................................................................... 1994, the figure was $1.89 billion. The average number of welfare recipients per month increased from 327,800 From St. Paul’s Hospital, Vancouver; BC. Canada, Depament of Emergency Medicine (GV, JMC); and the University of British Colum- in 1993 to 351,400 recipients in 1994. The 1995 estimates bia, Van-couver;BC, Canaah (GI! JMC), Centrefor Human Settlements for welfare payments are higher than in any previous year. (SMS). Income assistance is distributed to all beneficiaries in Received: February 27, 1996; revision received: July 3, 19%; accepted: BC on one particular Wednesday each month. It is a popJuly 16, 19%; updated: August 17. 1996. ular notion that drug and alcohol abuse escalate during Prior presentation: In part at the SAEM annual meeting, Denver; CO, the days after the distribution of income assistance checks. In the ED at St. Paul’s Hospital, a tertiary hospital located May 1996. in the heart of downtown Vancouver, BC. “Welfare Address for correspondence: James M. Christenson, MD, Department of Emergency Medicine, St. Paul’s Hospital, 1081 Burrard Street, Van- Wednesday” has gained notoriety of mythlcal proporcouver; BC V6Z IY6. Canada. Fax: 604-631-5488; e-mail: jim- tions. The ambulance crews that work in downtown VanChris@ unixg.ubc.ca couver on the day checks are distributed describe downI Health programs and social programs have common
overall goals and focus on the,same target populations, However, the policies of one are poorly coordinated with the policies of the other. It is possible that policies in one field can undermine achievements in the other.
~~
119
Welfare Payday, Verheul et al.
town Vancouver as “Mardi Gras” because of the large number of intoxicated people on the streets. Health, social service, and law enforcement workers offer anecdotal corroboration of an increase in health and social problems related to increased drug and alcohol abuse following check issue day. Such a phenomenon has been previously studied. Brunette et aL6 reported the association of increased daily volumes in EDs, jails, alcoholic receiving centers, and 911 calls with welfare check distribution in Minneapolis. This paper examines the association between monthly welfare payments and some broad measures of morbidity and mortality in Vancouver, BC. The social policy of monthly welfare payments on the last Wednesday of each month is examined from a health policy perspective of added burden to the system associated with the response to acute medical and social encounters in the 7 days commencing with welfare paydays. The null hypothesis is that there is no difference in the numbers of acute medical and social encounters by government agencies on the 7 days commencing with the monthly check issue day (Welfare Wednesday) compared with the 7 days commencing on other Wednesdays.
I ...................................................................... METHODS Study Design: We performed a retrospective analysis of multiple databases to determine the association of adverse social and medical consequences with welfare check distribution. The study was considered exempt from review by the institutional ethics committee. Setting and Population: Vancouver is a port city with a population of 460,000. The Greater Vancouver Regional District (GVRD) of BC consists of Vancouver and many clustered but distinct municipalities, with a total population of approximately 1,600,OOO. In 1993, 529,009 persons were recorded as receiving income assistance in the GVRD (excluding aboriginals and persons aged >65 years).’ This figure may double-count persons who discontinued and then resumed welfare during the year, thus overestimating the number of individual recipients. On average, 267,240 unique individuals per month received income assistance in all of British Columbia in 1992/1993,’ and of these, approximately 156,000 income assistance recipients reside in the Greater Vancouver Regional District. Nearly 1 person in 10 in the GVRD were in receipt of income assistance at any one time in 1993. ,
copy daily tally sheet and entered into a computer file, were used to explore daily comparisons. The data sheets as supplied were presumed to be correct. Data entry was manually checked for accuracy by a second investigator. The databases included: St. Paul’s ED visits, ambulance service calls, fire department calls, fire department medical calls, coroner-reported deaths, coroner-reported deaths due to drugs and alcohol, police arrests for public drunkenness, police records of assaults due to alcohol, police reports of assaults due to drugs, and police reports of domestic violence.
Data Analysis: Data were entered into a database for statistical analysis (SYSTAT 5.1, SYSTAT, Inc. Evanston, IL). We compared mean daily number of events on each of the 7 days of the week commencing with “Welfare Wednesday” and similar weekdays in other weeks. (e.g., comparing “Welfare Wednesday” and all other Wednesdays). We then compared the mean weekly number of events in the week commencing with “Welfare Wednesday” (Wednesday to the following Tuesday) with the mean number of events in weeks commencing with a Wednesday that was not a check issue day. “Events” are defined as the traditional unit of service for the agency: patients registering in the ED, responses in the fire department, and emergency responses in the ambulance service (excluding hospital transfers). For each database, the calculated means and SDs were compared using an analysis of variance and Scheffe’s post-hoc correction (daily comparisons) and a 2-tailed Student’s t-test (weekly comparisons). Due to these multiple comparisons, a p-value of
