Study sample
The study's retrospective design employed the concept of a natural experiment to address the problem of selection bias, the primary threat to validity in a nonexperimental study (
7,
8). Because the program's single team can provide only partial coverage of 911 calls identified as psychiatric emergencies, a comparison group of emergency 911 situations handled by regular police intervention was available. The study sample consisted of 73 psychiatric emergency cases handled by the mobile crisis team and 58 such cases handled by regular police procedures during the three-month period from October 1 to December 31, 1995.
Information was obtained from records of the public safety department and the mobile crisis program on subjects' demographic characteristics (age, race, and gender), homelessness status, whether the subject had a state psychiatric hospital admission in the previous six months, whether the crisis situation involved violence, the duration of police involvement, and the disposition for the crisis situation. Homelessness was defined as having no permanent residence and currently living on the street or in a homeless shelter.
Program measurements
Effectiveness. Program effectiveness was evaluated by measuring the differences between the hospitalization and arrest rates of the study groups.
Efficiency. According to Mayer (
9), achieving efficiency may be understood as "trying to achieve the most of a desired benefit in relation to a given level of expenditure. Efficiency is a way of choosing among alternative means for achieving a standard end." For psychiatric emergencies, the standard end, or common objective, of both mobile crisis programs and regular police services is to resolve or ameliorate the situation for the immediate protection of the health and safety of the persons involved and to facilitate access to any additional treatment or follow-up services needed.
Efficiency was evaluated by comparing the cost per case (cost per episode) for mobile crisis services and regular police services. The cost per case for mobile crisis services was calculated as total program costs and any psychiatric hospitalization costs divided by the number of cases served. Program costs included the department of public safety's contribution of one police officer per team per 7.5 hour shift plus the mental health center's expenditures associated with staffing and operating the program, including administrative and support overhead costs.
Cost per case for regular police intervention included the cost of police services per hour ($39.33) multiplied by the average amount of time required per intervention (1 hour and 51 minutes) plus any psychiatric hospitalization costs. For the purposes of this study, the cost per hour of police services included average salary and benefit costs plus public safety department overhead. That overhead included administrative and supervisory costs; dispatch, training, and other support function costs; and supplies, equipment, and other operating expenses.
Cost per case of psychiatric hospitalization equaled the total cost of residential treatment services used divided by the number of cases served. The average cost per case for psychiatric hospitalization was calculated by multiplying the average length of stay by the average daily rate for facilities used by the mobile crisis team and the police. Average daily rates and lengths of stay were used to neutralize the choice of facility in comparing costs.
Consumer and police satisfaction. Consumer satisfaction was evaluated in conjunction with routine follow-up services to persons served by the mobile crisis team during the three-month study period. This convenience sample of 32 individuals or families was asked to complete the eight-item Client Satisfaction Questionnaire (
10). Items are rated on a Likert scale of 1 to 4, with 4 indicating very satisfied and 1 very dissatisfied; possible total ratings range from 8 to 32. Consumer satisfaction was also evaluated by open-ended questions about whether needed services were received and what the consumer liked and disliked about services received.
Because the program is jointly operated by the community mental health agency and the county public safety department, police officers' satisfaction with the working relationship and the team's performance is a critical factor for the program's success. Police officers were asked to rate, on a survey designed by the author, their satisfaction with the performance of the mobile crisis team; the survey used a 5-point Likert scale, with 5 indicating very satisfied; 3, neutral; and 1, very dissatisfied. Also included were open-ended items for comments and suggestions.
The survey was distributed to officers present at each of three consecutive shift roll calls for immediate completion and return; 106 officers completed the survey.
Analysis. Chi square tests with Yates' correction for continuity and t tests with Levene's test for equality of variances were used to determine the statistical significance of differences between the two study groups. The probability level for statistical significance was set at .05.