U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics Special Report Deaths in Custody Reporting Program Mortality in Local Jails, 2000-2007 July 2010 NCJ 222988 ------------------------------------------------------------- This file is text only without graphics and many of the tables. A Zip archive of the tables in this report in spreadsheet format (.csv) and the full report including tables and graphics in .pdf format are available from: http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=2197 This report is associated with the Deaths in Custody Reporting Program. For a list of related publications go to http://bjs.ojp.usdoj.gov/content/dcrp/dictabs.cfm. ------------------------------------------------------------- Margaret Noonan BJS Statistician A total of 8,097 inmate deaths in local jails were reported to the Deaths in Custody Reporting Program (DCRP) from 2000 to 2007 (table 1). The program received data from at least 99% of the roughly 3,000 jail jurisdictions each year. Annually, more than 80% of the nation's jails reported no deaths in their custody. For the 8-year period between 2000-2007, 42% of the jails nationwide reported no deaths. The number of deaths in local jails increased each year, from 904 in 2000 to 1,102 in 2007. During the same period, the growth in the average daily population of jail inmates (31%) outpaced the increase in inmate deaths (22%). This resulted in a decline in the mortality rate of jail inmates from 151 deaths per 100,000 inmates to 141 per 100,000 between 2000 and 2007 (figure 1). (See Methodology for calculation of mortality rates.) ----------------------------------------------------------- Highlights * From 2000 through 2007, local jail administrators reported 8,097 inmate deaths in custody. Deaths in jails increased each year, from 904 in 2000 to 1,102 in 2007. * The mortality rate per 100,000 local jail inmates declined from 151 deaths per 100,000 inmates to 141 per 100,000 between 2000 and 2007, while the jail inmate population increased 31% from 597,226 to 782,595. * Annually, more than 80% of the nation's jails reported no deaths in their custody. * Deaths from any illness, including AIDS, accounted for more than half (53%) of all deaths in local jails. * Heart disease was the leading cause among illness deaths in local jails (42%). * Suicide was the single leading cause of unnatural deaths in local jails, accounting for 29% of all jail deaths between 2000 and 2007, but the suicide rate declined from 48 to 36 deaths per 100,000 inmates. * Between 2000 and 2007, the suicide rates were higher in small jails than large jails. In jails holding 50 or fewer inmates, the suicide rate was 167 per 100,000; in the largest jails, the suicide rate was 27 per 100,000 inmates. * After adjusting for differences associated with the age, sex, race, and Hispanic origin, suicide was the only cause of death that occurred at a higher rate in local jails than in the U.S. general population. ------------------------------------------------------------ From 2000 to 2007, between 3,072 and 2,924 jail jurisdictions, reported data to the DCRP (table 2). In 2000, nearly 9 out of 10 jails (86%) reported zero deaths. By 2007, 8 out of 10 jails (81%) of jails reported zero deaths. For the eight-year period between 2000 and 2007, about 8 in 10 jails (83%) reported zero deaths. For the same period, about 1 in 10 jails (12%) reported a single death, and less than 1 in 10 jails (5%) reported 2 or more deaths. During this time, no more than 2% of jails reported more than 3 deaths (appendix table 1). Between 2000 and 2007, suicides were the leading cause of death, but suicide rates continued to decline Suicide was the single leading cause of death in local jails (29%), followed by deaths associated with heart disease (22%), intoxication (7%), and AIDS-related causes (5%) (figure 2). No other single known cause of death, including accidents (3%) and homicides (2%), accounted for at least 5% of deaths. Collectively, deaths from any illness, such as heart disease, AIDS, cancer, and liver disease, accounted for the majority (53%) of jail deaths during 2000 to 2007. (See Methodology for a discussion on the nature of illness deaths). While suicide has been the leading cause of death in local jails since the 1980s, it has declined over time. The suicide rate among jail inmates was 129 per 100,000 inmates in 1983, but 10 years later the rate had declined by more than half to 54 per 100,000 inmates.***Footnote 1: See Suicide and Homicide in State Prisons and Local Jails, BJS Web, 21 August 2005.**** From 2000 to 2007, the suicide rate declined by about a quarter, from 48 to 36 suicide deaths per 100,000 jail inmates. (table 3). Besides suicide, AIDS-related deaths were the only other leading cause of death that declined among jail inmates. Over the 8-year period, mortality due to AIDS declined by 50%, from 10 to 5 deaths per 100,000 inmates. The mortality rates for deaths from all other leading causes of illness in jails during this time period showed little variation. Deaths due to accidents (4 per 100,000) and homicides (3 per 100,000) also remained relatively constant throughout 2000 to 2007. By comparison, the mortality rate for intoxication nearly doubled from 6 per 100,000 inmates in 2000 to 10 per 100,000 in 2007. Ten types of medical conditions accounted for about 8 in 10 illness deaths in jail Local jail administrators reported over 50 different medical conditions that led to an inmate's death between 2000 and 2007. Of these 50 fatal medical conditions, 10 accounted for 77% of all illness-related deaths in local jails (appendix table 4). Heart disease was the leading cause of illness-related deaths (42% of illness deaths), and was reported at least 3 times as frequently as any other death due to illness. Besides heart disease and AIDS, no single medical cause of death accounted for more than 5% of the illness-related deaths in jails between 2000 and 2007. Overall, 95% of deaths reported to DCRP during the 8-year period indicated that an autopsy was completed. A definitive cause of death could not be determined for 6% (or 489) of deaths in local jails, and an autopsy was conducted in the majority (93%) of these unresolved cases (not shown in a table). No additional information was provided in about two-thirds of all undetermined deaths. In the remaining cases, either the information given was insufficient to determine the manner of death or the autopsy was inconclusive. Over half of deaths occurred within 1 month of admission Nearly a quarter (24%) of deaths occurred within 2 days of admission (not shown in table). Overall, more than a third (38%) of deaths took place within the first 7 days (table 4). Over half (56%) of inmate deaths occurred within 30 days of admission. The distribution of inmate deaths occurring in the first month varied by cause. The majority of suicides (64%), intoxication (80%), and accidental deaths (62%) occurred in the first month. Deaths due to other illness-related causes were more frequently observed among inmates serving longer sentences. Over a third of cancer deaths (36%) and a fifth (21%) of AIDS-related deaths occurred more than 180 days after admission. The pattern of deaths since admission generally corresponds with the length of stay in jail and whether an inmate has been detained long enough to be at risk of dying during a certain time period. The number of deaths occurring within the first month of admission generally corresponds with the number of inmates who spend less than one month in jail. The Annual Survey of Jails (ASJ) found local jails admitted an estimated 13 million inmates during the 12 months ending June 29, 2007, with an estimated 80% (10.4 million) serving their time within one month. Data from the 2004 Survey of Large Jails (SLJ), which collected discharge data from the largest jail jurisdictions (1,000 or more inmates), found that more than two-thirds (46%) of jail inmates were released within two days of admission, and a majority (80%) served less than a month (figure 3). Between 3 and 6% of jail inmates served past day 31. While the survey was limited to the largest jails, these facilities held more than half of the U.S. jail inmate population (52%). These facilities' turnover rates (54%) were half the turnover rate for the smallest jails (110%). If the sample included all jails, the inverse relationship between expected length of stay and facility size would likely increase the frequency of inmates serving less than 30 days and reduce the pool of inmates at risk of dying during lengthier stays. Suicide rates in the nation's smallest jails were 6 times higher than in the largest facilities Suicide rates were higher in smaller facilities than in larger facilities. According to the 2005 Census of Jails, large jails holding an average daily population (ADP) of 1,000 or more inmates made up about 6% of all jails, and in 2007 these jails held 52% of the nation's jail population. About 40% of all jails held fewer than 50 inmates on an average day, and in 2007 these jails held about 3% of all jail inmates. The smallest jails had the highest average annual suicide rate (167 per 100,000 inmates), while the 50 largest jails had the lowest average annual suicide rate (27 per 100,000 inmates) (table 5). The lower rate of suicide in large jails may reflect the capacity of these jails to provide a variety of suicide prevention measures. According to the most recent data available in the 1999 Census of Jails, over half (54%) of jails holding fewer than 50 inmates provided staff training in suicide prevention, compared to 91% of the largest jails. Fewer than half (41%) of the smallest jails provided inmate counseling or psychiatric services, while such services were provided by over 90% of the largest facilities (table 6). Overall, the availability of various suicide prevention measures increased with facility size. This comparatively high suicide rate may be due in part to the high turnover rate and the resulting shorter inmate stays in small facilities.***Footnote 2: Turnover rate is calculated by taking the sum of admissions plus releases, and dividing by the average daily population. For more information on frequency of turnover, see Jail Inmates at Midyear 2007, BJS Web, 6 June 2008.*** The mortality rate for jails holding more than 50 but fewer than 500 inmates decreased as facility size increased. The mortality rate of facilities with an ADP of fewer than 50 inmates (282 per 100,000) was at least 1.5 times higher than that of any other facility size category. Medium-sized jails, holding between 100 and 499 inmates, had the lowest overall mortality rate (114 and 115 per 100,000). Larger jails (those holding more than 500 inmates) generally had higher mortality rates from illness deaths, such as AIDS, cancer, and liver disease. Small jails, holding fewer than 50 inmates, had an intoxication mortality rate (31 per 100,000) that was three times higher than the 50 largest jails (8 per 100,000). Also, the smallest jails had the highest accident mortality rate (11 per 100,000), and the highest rate of deaths without a known cause (13 per 100,000). However, homicide rates were highest in the 50 largest jails with 6 per 100,000 inmates, compared to 1 per 100,000 for the smallest jails. Despite the higher mortality rates experienced by the nation's smallest jails, these facilities accounted for a small percentage of all jail deaths (7%) (table 7). Jails holding at least 1,000 inmates accounted for 4,240 (52%) of deaths reported over the 8-year period. Between 2000 to 2007, with the exception of one jail in 2002, no jails holding fewer than 500 inmates reported more than 3 deaths. The 50 largest jail jurisdictions accounted for a third of all jail deaths in the United States The 50 largest jails accounted for about a third of the jail population in the United States and 35% of all jail deaths. These large urban jails accounted for about 60% of the jail deaths from cancer, AIDS, and homicide. In contrast, 19% of jail suicides and 23% of jail intoxication deaths occurred in the 50 largest jurisdictions. Heart disease was the leading cause of death in the 50 largest jail jurisdictions (41 per 100,000), followed by suicide (27 per 100,000). For all other jails, suicide was the leading cause of death, followed by heart disease. The death rate from AIDS in the 50 largest jails (15 per 100,000) was 7.5 times higher than the rate in jails holding less than 1,000 inmates (2 per 100,000) (not shown in a table). The homicide rate in the 50 largest jails was 3 times that of all other jails. The rate of intoxication deaths was 4 times higher in the smallest jails (31 per 100,000), compared to the 50 largest jails (8 per 100,000). ----------------------------------------------------------- Mortality differs among prison and jail populations Local jails and state prisons differ in their general purposes, populations housed, and terms of confinement.***Footnote 3:For details on the differences between prison and jail populations, see page 6 of Prison and Jail Inmates at Midyear 2000, BJS Web, 25 March 2001.**** Jails are locally operated facilities that confine persons before or after adjudication. In addition to convicted inmates, jails also hold inmates awaiting arraignment, trial, and conviction or sentencing. The annual volume of admissions into jails is about 17 times the size of the jail population at midyear. The mean time served in local jails is about 21 days. State prisons typically only house sentenced felons, with annual admissions that amount to just over half of their yearend population. In 2006 the mean time served by offenders released from state prison was 32 months. As a result of the differences between jails and state prisons, the number, rate, and pattern of deaths differed. Between 2001 and 2007, an average of 3,134 state prisoner deaths occurred each year, compared to 1,028 jail inmate deaths each year (table 8). The overall mortality rate in state prisons (251 per 100,000 prisoners) was 1.7 times higher than that of local jails (144 per 100,000 inmates) for the same period.***Footnote 4: The DCRP began collecting death data from state prisons starting in 2001. Prior to 2001, data was limited to counts of deaths by manner of death and sex. For more information, See Medical Causes of Death in State Prisons, 2001-2004, BJS Web, 21 January 2007.*** In both jails (22%) and state prisons (27%), heart disease accounted for about a quarter of all deaths. The rate of heart disease deaths in state prisons (69 per 100,000) was twice that reported by local jails (32 per 100,000). In state prisons, cancer also accounted for a quarter of all deaths (23%), while cancer accounted for 3% of jail fatalities. The cancer mortality rate in state prison (58 per 100,000) was 12 times that observed in jails (5 per 100,000). The suicide mortality rate in jails (42 per 100,000) was 2.6 times that in state prisons (16 per 100,000). ----------------------------------------------------------- Mortality rate for males was slightly higher than females Males were slightly more likely than females to die in local jails. The mortality rate for males was 146 per 100,000 compared to 136 per 100,000 for female jail inmates (table 9). Males (33 per 100,000) were 1.3 times more likely than females to die of heart disease (25 per 100,000), while females were 1.4 times more likely to die from all other illnesses (table 10) than males. Also, males (45 per 100,000) were 1.6 times more likely to commit suicide than females (28 per 100,000), and nearly half as likely as females to die from intoxication (9 and 17 per 100,000, respectively). The overall mortality rate increased dramatically with age, although older inmates accounted for a small proportion of the total population. Inmates age 55 and older accounted for 2% of jail inmates, but 14% of jail deaths. The mortality rate of inmates age 55 and older (893 per 100,000) was more than twice the rate for inmates age 45 to 54 (383 per 100,000), and 10 times higher than the rate for inmates age 25 to 34 (88 per 100,000), which was the largest age group of inmates.***Footnote 5: See Profile of Jail Inmates, 2002, BJS Web, 18 July 2004, and Jail Inmates at Midyear 2007, BJS Web, 6 June 2008.**** AIDS and intoxication were the only causes of death for which inmates age 55 and older did not die at the highest rate of all age groups. Inmates age 45 or older died at a rate nearly 5 times higher than younger inmates The overall mortality rate of inmates age 45 and older (475 per 100,000) was nearly 5 times that of younger inmates (98 per 100,000). For all age groups under 45 years old, suicide was the leading cause of death. Older inmates were 9 times more likely to die due to an illness. Specifically, older inmates were 11 times more likely to die from heart disease and 19 times more likely to die of cancer than inmates younger than 45. Among causes unrelated to illness, older inmates were between 1.5 times (suicide) and 2.9 times (accident) more likely to die than younger inmates. Hispanic inmates were less likely than whites or blacks to die in jail; blacks were less likely to commit suicide White inmates, accounting for nearly half of all jail inmates, died in local jails at a higher rate (172 per 100,000) than black (123 per 100,000) or Hispanic inmates (115 per 100,000). Whites (68 per 100,000) committed suicide at 4 times the rate of black inmates (16 per 100,000) and twice the rate of Hispanic inmates (34 per 100,000). Whites and blacks died of heart disease at comparable rates (34 and 35 per 100,000, respectively). Each group was nearly twice as likely as Hispanics (18 per 100,000) to die of heart disease. Heart disease was the leading cause of death for black inmates (35 per 100,000), and accounted for more than twice as many deaths as suicide among blacks (16 per 100,000). Blacks (13 per 100,000) were 4 times as likely to die from AIDS-related causes as white inmates (3 per 100,000) and about twice as likely as Hispanic inmates (6 per 100,000). Suicide was the single leading cause of death among Hispanic inmates (34 per 100,000). Violent offenders had the highest overall mortality, suicide and homicide rates Overall, violent offenders (191 per 100,000) and public order offenders (152 deaths per 100,000) were more likely to die in local jails than other offenders (table 11). Violent offenders (80 per 100,000) and public-order offenders (35 per 100,000) were the most likely to commit suicide. Specifically, offenders serving time for homicide, kidnapping, and rape (200, 195, and 194 per 100,000, respectively) had the highest suicide rates. Violent offenders also had the highest rates of cancer and homicide mortality. In particular, offenders serving time for homicide or kidnapping were the most likely to be killed (13 per 100,000) while in jail. Property (106 per 100,000) and drug offenders (102 per 100,000) had the lowest overall mortality rates. Drug offenders had the lowest mortality rates from suicide (19 per 100,000) and homicide (1 per 100,000). Public-order offenders had the second highest overall mortality rate (152 per 100,000), but the mortality rate for most categories of public-order offenders were below the overall average. The mortality rate for public order offenders varied from a high of 431 per 100,000 for probation/parole violators to a low of 61 per 100,000 for offenders held on weapons charges. The mortality rates for all other public-order offenders were equal to or less than 147 per 100,000. Leading causes of death in the jail population differed from the U.S. resident population Mortality in local jails differed from mortality in the U.S. resident population in both prevalence and leading causes of death. In the resident population, heart disease was the leading cause and suicide was the tenth leading cause of death between 2000 and 2006 (the most recent year finalized mortality data were available) (table 12). By comparison, suicide was the leading cause of jail deaths, followed by heart disease (table 13). Among both jail inmates and the U.S. resident population, the rates of heart disease deaths increased with age. Similar rates of death from heart disease were observed for persons under 45 years old in both groups. Among persons age 45 and older, the heart disease mortality rate in jails was higher than in the resident population. Cancer was the fifth known leading cause of death in jails, but the second leading cause of death in the resident population. Across all age groups, death rates from cancer in jails were consistently lower than in the resident population. The largest difference was observed among persons age 45 to 54. The U.S. resident population rate of cancer deaths (122 per 100,000 residents) in this age group was 7 times that of jail inmates (17 per 100,000 inmates). Jail suicide rates are consistently higher than suicide rates in the general U.S. population Suicide rates for all age groups were at least 3 times higher among local jail inmates than the general population. The largest differences were seen among persons age 55 and older. These jail inmates had a rate of suicide (69 per 100,000 inmates) that was 5 times higher than U.S. residents (14 per 100,000 residents). For violent deaths, such as suicide, specific data collected by the Centers for Disease Control and Prevention (CDC) allow for standardized comparisons between jail and U.S. resident mortality rates (table 14). Standardizing these rates accounts for variations in the age, race, and sex composition of the two groups. For example, males make up 88% of jail inmates, but 49% of the adult resident population. The jail population was adjusted through a direct standardization so that it would reflect the composition of the general population. For all groups, the adjusted jail suicide rates remain consistently higher than that of U.S. residents. The standardized jail suicide rate was 4.4 times higher than the suicide rate in the general population, and the standardized jail suicide rate was higher for both males and females. Males were 2.8 times more likely to commit suicide in jail. White males (126 per 100,000) had the highest standardized jail suicide rate, which was 5.7 times higher than the suicide rate of white males in the general population (22 per 100,000). The standardized suicide rate for Hispanic males was about three times as high as the general population. Black males in local jails had the lowest standardized suicide rate (20 per 100,000), which was about 1.5 times higher than in the general population (12 per 100,000 residents). The suicide rate of females in jail was five times that of the suicide rate of females in the general population Females account for 51% of the general population and 12% of the jail population, so adjusting the jail population to reflect the general population would deflate the standardized suicide rate for females. White females had the highest standardized jail suicide rate at 45 per 100,000 inmates. The standardized suicide rate for Hispanic females (21 per 100,000), but suicide among Hispanic females in jail occurred 10.5 times more than in the general population (2 per 100,000) post-standardization. Blacks had the lowest standardized suicide rate among females (7 per 100,000), but it was 3.5 times higher than the suicide rate of black females (2 per 100,000) in the general population. The actual suicide rate in jail varies from the standardized jail suicide rate based on race and sex Overall, the standardized jail suicide rate (57 per 100,000 inmates) was about about 1.5 times the actual (observed) suicide rate (42 per 100,000 inmates) in jails. Female suicide rates in jails are nearly equal regardless of the standardization, while the male suicide rate is 1.3 times higher post-standardization. With the exception of whites, the jail suicide rates both before and after standardization show little variation. Among females, the standardized suicide rate was lower for every race or ethnic group. Methodology Update to the Mortality in Local Jails, 2000-2007 Report As part of data processing review, BJS and the Census Bureau identified duplicate records by providing jail reporting units, the entities that submit records to BJS for the Deaths in Custody Reporting Program (DCRP), with a summary annual form at the close of a collection year to record the total number of deaths occurring within the jail facilities. BJS and Census Bureau staff compared the aggregate count of individual death records received to the annual summary number reported by a jail reporting unit. If discrepancies were noted, the reporting unit was contacted by either BJS or Census Bureau staff, and the discrepancy was resolved. After the annual accounting, the dataset containing the individual death records was again reviewed to identify duplicates via a visual inspection of names and dates. BJS's review of death records received between 2000 and 2007 initially excluded 51 deaths deemed to be duplicates or out of scope. A death was out of scope if the inmate was outside of the jail's jurisdiction when he/she died. Inmates on escape status, out on probation or parole, or on home-electronic monitoring were considered outside of the jail's jurisdiction for the purposes of this collection, and were therefore out of scope and excluded. A death was a duplicate if the death had either been reported by two different reporting units, one of which had custody of the inmate at the time of death or if the death had been reported to a different collection within the DCRP. Deaths occurring in the process of arrest were the most common example of a duplicate death reported to a different DCRP collection. Of these 51 exclusions, 34 records were duplicates and 17 were out of scope. Two jail jurisdictions contacted BJS after the report was released and asked for a review of their data. BJS used this opportunity to do a stringent assessment of the cases submitted by expanding the duplication identifying code to include cases that matched on state in which the death occurred, the gender of the inmate, and the date of birth or date of death of the inmate. Over a hundred potential duplicates were identified, thirteen of which were identified as true duplicates. The thirteen duplicate cases escaped detection during the earlier quality control process because the summary totals and number of death records within each reporting unit matched, and the information reported on names, admission date, charges against the deceased, and facility holding the inmate at death differed slightly between reporting units that submitted the same death records. Consequently, BJS efforts to match records on these variables in order to identify duplicate records did not flag these 13 records. The duplicate records reduced the total number of deaths occurring in local jails during the 2000 through 2007 period from 8,110 to 8,097. These duplicate records were spread across the eight-year period. In 2000, 2001, 2002, and 2006, one duplicate death was reported. In 2001 and 2005 two duplicate deaths were reported, and in 2004 five duplicate deaths were reported. No duplicate deaths were reported in 2003. These additional cases increased number of exclusions to a total of 64 deaths. BJS anticipates that there will be some level of duplication as we advise respondents to report deaths of uncertain eligibility with the assurance that BJS will exclude ineligible cases. DCRP mortality data are always based on information available to BJS at the time of publication. If addition information about a death occurring in custody comes to the attention of BJS, we incorporate updates into the data. As a result, findings may differ from earlier published reports. Deaths in Custody Reporting Program The local jail portion of the Deaths in Custody Reporting Program began in 2000 after the passage of the Deaths in Custody Reporting Act of 2000. The Governments Division of the U.S. Census Bureau, as an agent for BJS, collected facility-level information, death counts, and information surrounding the death from local jails. Data on inmate deaths were requested at the end of each calendar quarter. At yearend, jails submitted a reconciled annual count that summed to the quarterly count. Additionally, jail administrators submitted data on inmate population and admission counts, which were the basis for the calculation of inmate mortality rates. Jail respondents were instructed to report the death of any inmate in their custody, even if the inmate was held for other jurisdictions. Respondents were also instructed to include the death of any inmate in transit or sent outside the jail facility for medical, mental health, or substance abuse treatment services. Inmates who died while on or traveling to or from work-release programs were also included. Deaths of jail inmates who were released on temporary furloughs or who had escaped the jail facility were excluded. Medical cause of death information was collected for illness deaths, but not for other manners of death. The illness category is a heterogeneous category for natural causes of death associated with an underlying illness. Due to the variation within the illness category, each illness death was discussed in terms of the known medical cause of death as determined by the autopsy. For each death marked as illness/natural cause, respondents were directed to specify a medical cause of death based on an autopsy review, if available. These text entries were later coded by clinical data specialists according to the World Health Organization's International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10). BJS analysis of causes of death used categories of disease published by the National Center for Health Statistics (NCHS). For this report, accidents, suicides, and homicides were discussed in terms of cause of death. While there is a clear distinction between manner and cause of death from a medical-legal standpoint, no such distinction was made in this report. More information on state prison death data and the Deaths in Custody Reporting Program is available on the BJS Website at . Average daily population used in calculating mortality rates Mortality rates were calculated as the number of deaths per year divided by the annual average daily population (ADP) of jail inmates and expressed in terms of deaths per 100,000 jail inmates. The measure of ADP was for the 12-month period ending December 31. The ADP implicitly measures the number of jail inmate days per year, thereby annualizing the mortality rates. The ADP reflects the number of admissions and their length of stay and can be expressed as the product of the number of admissions times mean length of stay. When mean length of stay is expressed in years, the ADP is equivalent to the number of person-years spent by jail inmates during a given year. This allows for reporting annualized mortality rates that are comparable in concept to the annual mortality rates reported by the National Center on Health Statistics. NCHS calculates (crude) mortality rates as the number of events for a period (e.g., year) divided by the population estimated at the midpoint of the period. For general population mortality statistics the use of the population at the midpoint of a year is employed as an approximation to the average population exposed to risk of death during the year. (See, e.g., Jacob S. Siegal and David Swanson (2004). The Methods and Materials of Demography, Second Edition, San Diego, CA: Elsevier Academic Press, p. 269.) Because length of stay in jail is comparatively short (i.e., about 21 days) the ADP is preferred over the midyear jail population count in the estimation of the average population exposed to risk of dying in jail. The average daily population is an appropriate measure because it is the closest measure of the accepted epidemiology model of "person years," or the number of years the person lives in the population prior to death (Gaswirth, J.L. (2006) "Case Comment-Buncher v. Brown County: the need for an appropriate comparison of suicide rates." Law, Probability and Risk (2005) 4, 257-263). Unless otherwise specified, the jail mortality rate was expressed in terms of the number of jail inmate days throughout the report. For 2000 and 2001, data on ADP were not collected by the DCRP. To estimate each jail's ADP for these two years, an average of the inmate custody population on January 1st of the reference year and on December 31st of the year was used. From 2002-forward, the ADP was collected by the DCRP and was as the estimate of the population at risk of dying in jail. To estimate the ADP of the characteristics of jail inmates-such as age, race, and gender-population proportions from BJS' 2002 Survey of Inmates in Local Jails were generated and applied to the average of each year's ADP. These were the most recently available data for selected jail inmate population characteristics nationwide. Data from BJS' Annual Survey of Jails indicate that the distribution of inmate characteristics such as gender and race have changed slowly over time. See Jail Inmates at Midyear 2008 - Statistical Tables, BJS Web, 31 March 2009 for more information. Using the 2002 inmate survey data to estimate the distributions of inmate characteristics and applying these distributions to the average ADP over the period results in a reasonable estimate of the ADP of inmate characteristics. U.S. resident population sources The NCHS's vital statistics program provided the U.S. resident population mortality data and rates presented in this report. See the NCHS website for mmortality data for the U.S. resident population at . The Center for Disease Control's (CDC) Web-based Injury Statistics Query and Reporting System was the source for age-specific suicide data in the U.S. resident population. See for more information. Medical cause of death in state prisons 2004 was the last year finalized clinical ICD data was available for state prisons. ICD distributions from 2001-2004 were applied to prison data reported to the DCRP for 2005-2007. More information on state prison death data and the Deaths in Custody Reporting Program is available on the BJS Website at . Discharge data Inmates are typically discharged from jail through bail, release by a court order, community supervision, time served or death. Estimates of time served by inmates discharged from jail are based on unpublished BJS data. Direct standardization of suicide rates in jail and U.S. resident populations The method of direct standardization was used to compare suicide rates in the jail inmate population with suicide rates in the general population. Because the characteristics of the jail inmate population differ from the characteristics of the general population, and as some characteristics are associated with suicide rates, a direct comparison of the crude suicide mortality rates between jail inmates and the general population can be misleading. For example, in jails, older white males are more likely to commit suicide than other groups, but older white males comprise a smaller fraction of the jail population than they comprise of the general population. In order to compare the jail and U.S. resident population suicide rates, the differences in these two populations' gender, race, and age distributions were taken into account by method of direct standardization. The jail population was adjusted to the age distribution of the general population, creating the standardized population; this was done within race and gender groups. The age-specific jail suicide rates were applied to the standardized population to produce an estimate of the expected number of suicides that would occur in jail if the age distribution of the jail population were the same as that of the general population. Again, this was done within race and gender groups. The expected numbers of suicides across age groups were summed and the total divided by the ADP of jail population to produce the standardized suicide rates that were compared to the general population rates. ------------------------------------------------------------- The Bureau of Justice Statistics is the statistical agency of the U.S. Department of Justice. James P. Lynch is director. This Special Report was written by Margaret Noonan. William Sabol and Spencer Li verified the report. Doris J. James, Catherine Bird, and Jill Duncan edited the report, Barbara Quinn and Tina Dorsey produced the report, and Jayne E. Robinson prepared the report for final printing under the supervision of Doris J. James. July 2010, NCJ 222988 ------------------------------------------------------------ This report in portable document format and in ASCII and its related statistical data and tables are available at the BJS World Wide Web Internet site: http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=2197. ------------------------------------------------------------ Office of Justice Programs Innovation -Partnerships- Safer Neighborhoods http://www.ojp.usdoj.gov ------------------------------------------------------------ 7/6/2010/JER/9:00 /Edits/TD/10:10/ Revised/TD/10/26/2010