REFERRAL REDESIGN

Partnering with Vendors

Since 1994, this plan tracked issues related to its members' satisfaction with the referral process. It had measured complaints, appeals and member satisfaction. Seeking to improve its members' understanding of the referral process, it improved its communications regarding urgent and emergent care, as well as the referral process in general Yet it found that even though member understanding of the referral process was improving over time, their satisfaction with it was not.

The number one member appeal issue was "noncovered emergency room." Twenty-six percent of all member appeals revolved around this issue.

The second most common cause of appeals was "nonemergency referral." Twenty-four percent of all appeals concerned this issue. The plan targeted these top two issues for improvement, and began a redesign of the referral process. By improving the administration of referrals, the plan hoped to boost satisfaction with the process.

PARETO ANALYSIS There is a saying in medicine: "Common things happen commonly." This is as true of service issues affecting member satisfaction as it is of clinical issues affecting outcomes. In this case, two issues accounted for half of all member appeals. The plan focused its efforts by looking at these two issues. In more complex cases, a simple histogram of the frequency of the "top ten" causes for a problem can help prioritize a response. This technique is named for Wilfredo Pareto, a nineteenth century economist and sociologist. The Pareto Principle states that a small number of causes are responsible for a large percentage of the effect-usually a 20-percent to 80-percent ratio.

The plan decided to address three aspects of the referral process: member satisfaction, provider satisfaction and administrative services.

The plan targeted its entire enrolled population for this activity. It used five performance measures:

• Percent of members satisfied with the referral process as measured by the plan's quarterly member satisfaction survey. The plan used an outside vendor for this survey. A random sample of members over age 18 was selected for the survey mailing. Over 500 surveys were received; 436 answered the salient question, "How satisfied are you with the process of getting a referral to a specialist?" A response of "satisfied" or "very satisfied" was included in the percentage reported. The plan restricted the survey to members enrolled for at least six months. It ensures that no member is mailed a survey more than once a year
• Percent of members satisfied with the referral process, as measured by the annual HEDIS satisfaction survey question: "How much of a problem was not being able to get a referral to a specialist that you wanted to see?" A response of "not a problem" was regarded as an indicator of satisfaction. This measure, added to HEDIS in 1996, was used as an adjunct to the plan's own quarterly surveys
• Percent of providers satisfied with referral policies, as measured by the plan's annual PCP satisfaction survey. This survey was administered to a random sample of offices. The sample was divided equally between physicians and office managers. A total of 411 surveys evaluated the referral process in the baseline year
• Appeals per 1,000 members. This was measured annually using appeals and grievance reports. The plan also performed a drill-down analysis, looking at the number of appeals related to referrals and emergency room claims
• Actual number of member complaints regarding referrals received by member services

Baseline performance was measured in 1995. Examining baseline performance, the plan set yearly goals for the first measure (percent of members satisfied with referrals) and the fourth measure (appeal rate per 1,000 members). It used a benchmark performance from a competing plan to set a 1998 goal for the third measure (percent of providers satisfied with referral policies).

The plan used multiple regression analysis to determine that satisfaction with the referral process was a significant predictor of overall satisfaction for both members and providers. Drill-down analysis revealed that the most important factor in member dissatisfaction was the process of getting a referral to a specialist, and that the leading reason for member appeals was rejection of emergency room claims.

Based on the results of the drill-down analysis, the plan decided to focus its efforts on removing unnecessary barriers to emergency care and streamlining referrals to specialists.

REGRESSION ANALYSIS Most plans carrying out quality improvement activities use only basic statistical techniques, such as the chi square or t test, to decide whether the improvements they observe after an intervention would be likely to occur by chance, or are probably due to the actions they have taken. This plan used a more advanced statistical technique-multiple regression analysis-to look at the strength of the relationships between several factors and overall satisfaction. The analysis showed them that satisfaction with the referral process was a major contributor to overall satisfaction for both providers and members.

To ease the use of emergency services, the plan did the following:

• Clinicians identified several "open wound" diagnoses that are always emergencies, and these were added to the "always pay" category of emergency room claims
• Requirements for authorization before emergency ambulance services, and for referrals for follow-up visits for out-of-area urgent and emergent care were removed, since these services are rarely used in nonemergency situations
• Member Services was given expanded authority to approve ER claims that had been previously denied because of lack of a referral
• A number of written guidelines were developed to help Member Services personnel resolve ER claims more quickly

To streamline the specialist referral process, the plan took the following steps:

• It removed the requirement for a referral before oncology care
• It expanded PCPs' ability to enter retroactive referrals from five days after the specialty visit to 180 days after the specialty visit

The plan also implemented outreach and education programs concerning emergency care and referrals, aimed at both members and providers:

• Member newsletter articles focused on the referral process, urgent and emergent care
• Member services staff were given specific instructions for explaining the referral process to members
• A 24-hour, toll-free telephone service, staffed by registered nurses, was started to help members with access to care questions
• A brochure entitled, "The Referral Process is Easier Than You May Think," was distributed to physicians' offices

The first remeasurement was based on 1996 data:

Survey data showed mixed results. The first measure (of member satisfaction) showed improvement, but was statistically equivalent to baseline. 1996 was the first year of HEDIS survey data; it served as a baseline for future years. Provider satisfaction with referral policies showed statistically significant improvement.

The appeals rate was reduced by nearly half, well ahead of the plan's goal-a statistically significant improvement. This was driven by reductions in referral appeals (from 1,181 in 1995 to 794 in 1996) and emergency room claims appeals (from 1,287 in 1995 to 505 in 1996).

DURING REENGINEERING In 1996, the Member Services department underwent extensive reorganization as part of this reengineering effort. The plan centralized member services functions, shifted accountability for reporting, and changed reporting methodology. All of these changes led to partial year reporting on member complaints in an inconsistent format. Although subsequent years saw the return of data reliability, 1996 data on complaints were essentially lost. It's been said that it's difficult to work on a car when the engine is running, but this is the essence of reengineering work processes in an organization with ongoing operations. Although some tasks may end up being neglected, conscious management effort is required to ensure that "mission critical" tasks continue to be performed.

With substantial improvement in emergency room issues in the first year of this initiative, the plan prioritized nonemergency referrals for its next set of interventions.

The plan made several changes in the referral process:

• It extended the referral submission window for providers from five days to 180 days
• It implemented improved electronic tools for providers to manage referrals
• Member service personnel were provided with new tools outlining how to resolve referral problems quickly. They received training regarding the process members needed to follow when seeking out-of-area urgent care
• A more "member friendly" waiver form (used when members self-referred and took responsibility for payment) was introduced to specialists

The plan continued its efforts to educate members and providers on the referral process:

• It developed and distributed to PCPs a template letter for welcoming new members to their panels. The letter contained information about how the referral process worked
• It further promoted the toll-free, 24-hour line for access questions by distributing a wallet card to members

In addition, it convened a work group to consider additional improvements to the emergency room claims process. At two hospitals, it piloted a system of e-mailing PCPs when members used the emergency room without referrals.

The second remeasurement used data from 1997:

The first four measures showed significant improvement over the previous year; the fifth measure improved significantly over baseline. (No results had been reported for the fifth measure in 1996.) The appeals rate continued to be driven downward by reductions in referral appeals (from 794 in 1996 to 252 in 1997) and emergency room claims appeals (from 505 in 1996 to 137 in 1997).

Although gratified by its performance, the plan still felt that there was room for further improvement in the referral system. It focused on a number of new enhancements.

It worked with its mental health vendor to improve referral processing for PCP-referred psychiatric ER visits.

It introduced a new standardized approach to problem solving by plan staff. Over 700 employees were trained in this process, which emphasized reasonable and flexible decision authority.

It implemented a five-part plan to revamp the referral system:

• The maximum number of visits covered by a referral was expanded from three to 99 and the length of time for which a referral was valid was expanded from 90 to 365 days
• The requirement for an authorization was removed from a number of services -this allowed PCPs to more fully control their referrals, with less "double-checking" by the plan
• It eliminated its production of waiver forms for self-referred patients
• It decided to pay for urgent care rendered in an emergency room setting
• It continued member education, targeting members who switched PCPs

The plan included training for its member services personnel, and notification of its providers as parts of its implementation process for the new referral system.

Remeasurement number three used 1998 data:

Both member survey scores were statistically equivalent to results from the previous year, but the provider satisfaction score showed significant improvement over the prior year. Appeals showed a significant decrease, once again being driven by reductions in referral appeals (from 252 in 1997 to 97 in 1998) and emergency room claims appeals (from 137 in 1997 to 103 in 1998). Complaints showed a similar decrease.

The plan has continued all of its activities, showing sustained levels of satisfaction in 1999. The year 2000 brought changes in a variety of state laws regarding appeals (prudent layperson standard for ER coverage, continuity of care provisions and direct access to obstetricians). The plan recognizes that this may result in some comparability issues for prior years data. Its long record of incremental improvement in patient satisfaction leaves it well poised to meet the demands of the new century.