Regional Shared Staffing as a Long-Term Strategy to Manage Staffing Crises

Recruitment and retention of hospital medicine clinicians have long been major challenges for healthcare organizations nationwide. Limited research and education opportunities, work, and personal considerations have made recruitment and retention of board-certified hospitalists to smaller community-based practices an additional challenge.

More than 120 physicians and advanced practice providers (APPs) of the Mayo Clinic Midwest Community Hospital internal medicine practice (of which 90% are physicians) provide 24/7 hospital medicine services to the Rochester, Minnesota, community and across the Mayo Clinic Health System in Minnesota, Wisconsin, and Iowa. The ongoing staffing deficits, especially in the Southeast Minnesota (SEMN) region, force the practice to rely heavily on locum tenens staffing to supplement the 24/7 coverage and address variations in patient volume and acuity level across the region.

Locum tenens staffing, while addressing the physician shortage, offers unique practice challenges. Multiple studies(1,2) and practice experience have shown that locum tenens staff, in addition to being less familiar with the healthcare system and served patient populations, are less concerned with readmission rates, patient satisfaction, and quality of care than hospital staff, regardless of how long they have been with the organization.

In late 2017, a multi-disciplinary team represented by local, regional, and division key stakeholders including hospitalists, nurse practitioners (NPs), physician assistants (PAs), operations administration, as well as functional support from areas such as management engineering, human resources, business analysis, credentialing, and quality data and analysis, was formed to recalibrate the SEMN hospital medicine staffing to be in line with the overall Mayo Clinic Midwest optimization plans and reflect care team models to retain and care for appropriate patients locally.

Guiding principles for optimization were:

1. Meet staffing needs and realize the financial benefits of creating a sustainable long-term plan while maintaining a high level of patient care.(3)

2. Learn from the successful NP/PA incorporated hospital medicine model in two Mayo Clinic Community Hospital Medicine sites and external benchmarking.(4-6) The 2018 State of Hospital Medicine report conducted by the Society of Hospital Medicine revealed that “Use of NPs/PAs by academic hospitalist groups is up, from 52.1% in 2016 to 75.7% in 2018. For adult-only groups, 76.8% had NPs/PAs, with higher rates in hospitals and health systems and lower rates in the West region.”(7)

3. Attend to patient safety, appropriate workload, and acuity level of patients seen by consultants and APPs; ensure new hires get appropriate and sufficient training.(8)

4. Standardize clinical knowledge across the practice, including knowledge guided by approved protocols and used consistently across all sites.

5. Create a positive care-team environment between providers and staff through day-to-day collaboration. This improves provider and staff satisfaction, which subsequently reduces burnout and turnover rate.

The proposed models are expected to be phased in over three years. The new models would provide adequate staffing at a cost savings of more than $2.18 million per year once fully implemented. This will align hospital performance with the overall financial goals of the organization.

Assessment of the Current State

Of the four small community hospitals and two critical access hospitals (CAHs) in the SEMN region, Mayo Clinic Health System owns five; the other is owned by Allina Health.

The regional hospitalist staffing model started with 24.04 MDs and 3.8 APPs at the cost of $9.1 million in average salary and benefits. The smaller community hospitals are staffed day and night by in-house physicians, and the two CAHs are staffed during the day by physicians who take overnight calls at home. The average daytime census for the four community hospitals is 15–18 patients with an average daily admission of five, based on the organization’s 2017 data. The CAHs have an average census of eight patients with one to two admissions a day. The average annual staff retention rate (excluding retirements) is around 87.2% for the entire region.

The staffing deficit is addressed with a combination of physicians picking up extra shifts, moonlighting fellows, and supplemental staff; 6% of shifts are filled by locum tenens providers at the cost of $1.4 million (see Table 1).

Methods

A systematic approach was used to understand the status quo, determine the contributing causes of the high locum tenens use in the region, and assess potential interventions and readiness to narrow, if not close, the gaps (see Figure 1).

Figure 1. Systematic approach employed by the team

The team conducted a series of key stakeholder interviews and analyses to understand the respective needs, issues, and requirements of each key player (see Table 2). Several shadowing sessions were offered so the team could see firsthand the various models and how each works.

Based on analysis of existing models and information collected, the team conducted a cause-effect analysis to identify reasons for the high locum tenens use in the region (see Figure 2). Four major factors (staff, culture, siloed structure, and environment) were identified and (4 < 10) 10 possible root causes were identified (see Table 3). Each cause was assessed based on the estimated effort and impact of addressing each of the gaps (see Figure 3).

Figure 2. Root cause analysis identified potential causes of high locum tenens utilization/costs

Figure 3. Impact/effort grid to prioritize all identified causes and determine those the team can have an impact on

The team identified the following four “potential home runs” that have substantial buy-in from all stakeholders and for which resources were available at the time to make the most impact:

1. NP/PA underutilization;

2. Reactive versus proactive recruiting strategy;

3. Decentralized recruitment and scheduling process; and

4. Absence of a multi-site privileging and credentialing process, which prevented staff from working across all sites.

While a comprehensive picture of the practice’s current state was being created, a concurrent effort was underway to gain an independent perspective of other successful hospitalist staffing model options and/or best practices in the industry.

Research yielded six models with some fairly novel models for the team to consider(9) (see Table 4):

1. Classic paired rounding/dyad model;

2. Evolved dyad model;

3. Late afternoon and evening admission;

4. In observation unit or with lower acuity observation patients;

5. In cross-coverage and triage roles; and

6. Co-management services.

The team also benchmarked nine medium- to large-sized external healthcare systems.

The findings show that among all nine healthcare systems interviewed, eight responded that they used APPs in their hospitalist models. One was in the process of updating its hospitalist model to move away from using all MDs and toward an MD/APP integrated model.

Among the eight systems with integrated NP/PA hospitalist staffing models, 38% allow their APPs to function autonomously and conduct initial and ongoing assessments/examinations, develop treatment plans, write progress notes, place orders, and communicate with primary care providers and consultants. These healthcare systems also cited the important leadership roles APPs assume, such as president of the hospital’s hospitalist program, member of the hospital’s peer review and credentialing committee, or team leader of quality improvement activities. Another 25% assign their APPs to medically stable patients, primarily focused on ensuring that patients’ treatment plans are proceeding well and adjusting most medications while MDs do all of the admission histories, physicals, and consults, and most of the discharge summaries. About 13% of NP/PAs assume a triage role and rotate carrying the triage pager every day (see Figure 4).

Figure 4. Research and benchmarking findings

Analysis, Improvement, and Implementation

A multi-faceted operational plan addressed the identified high-priority areas and clearly outlined roles and responsibilities.

Brainstorming and evaluation resulted in the selection of three key interventions as the most effective and efficient way to address the four “potential home runs.”

1. Integrated NP/PA staffing model

A study of existing models suggested that an integrated NP/PA staffing model can support the regional shared staffing model that is in alignment with the overall Midwest strategy and would support the practice’s long-term sustainable development.

Based on the resource capabilities and the readiness of each site, the new models were piloted, studied, modified, and improved in selected sites before a full-scale rollout.

The results of the pilot showed a 14% increase in average daily hospital census and an 8% decrease in hospital patient transfer to another facility. Additionally, when the new model was implemented at the pilot site where the providers work in both the outpatient and inpatient setting, the providers were able to go back to the outpatient clinic setting, which allowed for more appointment availability, as well as to be alleviated from having to keep up their skills in hospital internal medicine.

With the successful pilots, the dyad model was selected as the base model to be adopted and modified to best accommodate the local population and resource capabilities. After a standardized NP/PA training and structured orientation of 3–6 months, depending on the APP’s capabilities and experience, combined with a boot camp, online didactic modules, competency evaluation, and ICU in-service training, the APPs function relatively autonomously across the region and manage their own patients; the physician is available for backup/consultation as needed.

The new model was rolled out to staff the entire region with 16.00 MDs and 15.30 NPs/PAs, which provides adequate staffing to take care of the patient acuity and census across the region. Sites’ day shifts are staffed with board-certified MD consultants and well-trained NPs/PAs; night shifts are staffed primarily with NPs/PAs.

Post-implementation feedback received from both MD and NP/PA groups revealed that NP/PAs expressed consistent support from MDs for patient care; the overall census for physicians decreased from 15–18 to 12–15 patients per physician, with NPs/PAs independently seeing 6–8 patients per APP; physicians’ night shift burden decreased significantly as night shifts are primarily covered by NPs/PAs.

2. Centralized recruitment and scheduling model

In the reactive recruitment model, the practice typically waited for a position to open up and then hire for the position. Although the pull-based strategy may work in major medical centers, for smaller community hospitals like the ones in the SEMN region, this reactive hiring strategy creates a lag-time of at least 6–9 months before a hire is added to the work schedule. It could take even longer if there were internal or external competitions for similar top-tier talent at the same time.

With the new staffing model identified, the practice is now able to deploy strategies like centralized recruitment to proactively hire for positions that need to be filled.

Several teams looking to address similar improvement opportunities within the enterprise collaborated with the Human Resources and Recruitment Department to develop a centralized recruitment model that has a single recruiter for the entire Midwest hospital medicine practice. This eliminates internal competition and the delays caused by process variations and seeks to place the best hospital internal medicine providers.

The practice also moved to a single, integrated provider scheduling software to create a standard block schedule with a Wednesday shift-week start date for all HIM staff as a centralized scheduling strategy.

3. Multi-site privileging and credentialing process

The Mayo Clinic’s Midwest credentialing and privileging processes have traditionally been based at the site level or regional level in various forms. The new staffing strategies, along with increased practice integration within the health system, challenged the credentialing and privileging staff and supporting processes and drove the need for more cross-credentialing at the health system level.

In addition to the state regulatory differences, credentialing requirements vary across each site relative to practice/medical staff requirements, medical staff bylaws and medical staff policies, and committee approval processes and timelines.

A high-level process map (see Figure 5) was developed for the MCHS credentialing and privileging process, in addition to detailed swim lane maps of each region’s credentialing and privileging processes and committee approval structure.

Figure 5. Current high-level credentialing and privileging process

The high-level process can be summarized into four main steps:

1. Initiate the process through a request or notification.

2. Collect relevant information from physicians/scientists/advanced practitioners.

3. Verify the information with primary sources.

4. Approve credential verifications and grant initial hospital privileges.

Based on analysis of the existing processes, wastes and variabilities were identified across all of these major steps, including waiting/delay, under-used talent, and extra/inefficient processes (see Figure 6).

Figure 6. Wastes and variabilities identified across four major steps of credentialing and privileging process

A Midwest Credential Governing Council will coordinate and facilitate the integration of common processes for collecting and verifying provider information across the MCHS. Membership will include physician partners and departments/offices currently providing credential and privilege services. The reporting structure is recommended through the Mayo Midwest Personnel Committee.

The new centralized privileging and credentialing model has a single “point of entry” for information collection from the providers to reduce the number of offices and functional areas contacting providers to collect the same information for privileges at multiple sites. All provider credentialing information will be stored in a single common repository. A single point of contact was established for providers and practice leaders to track status as well as post any process-related questions. The approval bodies and processes were also consolidated and simplified to accelerate the process (see Figure 7). The new centralized model decreased the end-to-end process time, from requesting hospital privileges to privileges being granted, from more than six months to a few weeks.

Figure 7. Proposed credentialing and privileging process

The centralized privileging and credentialing process smooths the sharing of staff among sites to manage census and acuity variation across the region and allows all HIM staff to choose whether they are willing to cover shifts at any other SEMN sites or regions within MCHS (to preserve and enhance education and provider expertise caring for different patient populations/acuities), or work in the home base as they preferred, without going through the privileging process multiple times.

Results

By the end of the first year of implementation, the percentage of shifts covered by locum tenens staff decreased from 6% (sample size of 54,080 hours) in 2017 to less than 1% (sample size of 51,391 hours) in 2018, at a cost savings of more than $1.29 million in labor (see Table 5). In 2020, three years after the new models were in place, despite an organizational staffing change in the context of COVID-19, the practice reports no locum tenens use in the year.

Further incremental margin analysis was conducted to validate the financial viability of the new model (see Table 6). Benefits of an integrated APP model as revealed by the optimization effort include:

1. Free up time for MDs to spend with patients by significantly reducing their “legwork and paperwork.”

2. Ensure continuity of care provided by a care team member.

3. Enhance cost-efficiency and provider productivity.

4. Orient new physicians to the hospital.

5. Act as liaisons between the hospitalists and other providers.

6. See patients more promptly for initial assessments.

7. Spend more time communicating/educating patients and family members.

8. Bring a second set of eyes and ears to patients with complicated cases.

In reducing locum tenens cost and use, the team did not want to increase the Case Mix Index-Adjusted Average Length of Stay (CMI-Adjusted ALOS), which is defined as the ratio of the number of days of hospital care that were used to care for patients adjusted for the documented severity of the illnesses. The counterbalance measure was to ensure that the efficiency of patient-centered care is managed while the average severity level of cases didn’t change throughout the implementation. Research shows that a good reference CMI-Adjusted ALOS is 2.75 days or below.(10) And the practice’s CMI-Adjusted ALOS actually decreased from 2.8 days in 2017 to 2.5 days one year post-implementation (see Table 7).

Other measurements such as hospital patient transfer to another facility also show favorable outcomes: a significant decrease from 21%–37% to 0.5%–5.1%, depending on site.

Although the average annual staff retention rate (excluding retirements) didn’t decrease notably in the first year of implementation, it began to show improvement in 2019 at 94.4% for the region.

A control plan to further monitor and sustain the change was put in place for HIM Practice leaders and local operational leaders to meet and review the metrics routinely and take corrective actions if needed. If the percentage of shifts or hours covered by locum tenems exceeds the 2% threshold at any time, the chair and operation administrator of the division of Midwest Community Hospital Medicine will bring it to the next division meeting to examine the change in staffing and determine a detailed action plan.

Conclusion

Every crisis is an opportunity to address the root causes and find a solution that not only meets the current crisis, but also potentially avoids future crises. With the current relaxation of hospital bylaw restrictions on APPs due to the COVID-19 pandemic, the role of APPs has become an even more crucial lever to finding more cost-effective care delivery models.(11)

Creating change in the practice takes time, continuous improvement, and considerable support from the leadership to ensure successful development, implementation, and sustainability. Without a shared vision, full representation, and early engagement from all affected stakeholders, leadership’s commitment and continuous support on optimizing and sustaining the practice, or an effective change management plan, crisis management could easily turn into responsive disaster recovery planning.

To create a cost-efficient, sustainable regional hospital medicine model and develop a longer-term plan of hospitalist need, considering retirements, average tenure, and strategy to share staff among sites, the Mayo Clinic Midwest Community Hospital Medicine Practice has identified the integrated NP/PA Hospitalist staffing model as an effective strategy to manage long-term staffing crisis. Once fully implemented, the evolutional staffing models aim to provide more than $2.18 million per year in cost savings and eliminate the reliance on external providers to care for our patients.

References

1. Quinn R. The Pros and Cons of Locum Tenens for Hospitalists. The Hospitalist. December; 2012(12). www.the-hospitalist.org/hospitalist/article/124988/pros-and-cons-locum-tenens-hospitalists . Accessed November 28, 2017.

2. Ferguson J, Walshe K. The Quality and Safety of Locum Doctors: A Narrative Review. J R Soc Med. 2019; 112(11):462–71. doi:10.1177/0141076819877539

3. Cardin T. NPs, PAs Vital to Hospital Medicine. The Hospitalist. March; 2106(3). www.the-hospitalist.org/hospitalist/article/121858/nps-pas-vital-hospital-medicine . Accessed December 13, 2017.

4. Nelson J. Effective Hospitalist Roles for NPs, PAs. The Hospitalist. December 16, 2016. www.the-hospitalist.org/hospitalist/article/120379/practice-management/effective-hospitalist-roles-nps-pas . Accessed December 13, 2017.

5. Kleinpell RM, Grabenkort WR, Kapu AN, Constantine R, Sicoutris C. Nurse Practitioners and Physician Assistants in Acute and Critical Care: A Concise Review of the Literature and Data 2008-2018. Crit Care Med. 2019 Oct; 47(10):1442–49. doi: 10.1097/CCM.0000000000003925. PMID: 31414993; PMCID: PMC6750122.

6. Morgan PA, Smith VA, Berkowitz TSZ, Edelman D, Van Houtven CH, Woolson SL, Hendrix CC, Everett CM, White BS, Jackson GL. Impact of Physicians, Nurse Practitioners, and Physician Assistants on Utilization and Costs for Complex Patients. Health Aff (Millwood). 2019 Jun;38(6):1028–36. doi: 10.1377/hlthaff.2019.00014. PMID: 31158006.

7. The State of Hospital Medicine in 2018. The Hospitalist. January 8, 2019. www.the-hospitalist.org/hospitalist/article/192292/mixed-topics/state-hospital-medicine-2018/page/0/3 . Accessed May 20, 2019.

8. Designing an Effective Onboarding Program. The Hospitalist. March 19, 2020. www.the-hospitalist.org/hospitalist/article/219292/leadership-training/designing-effective-onboarding-program/page/0/1 . Accessed August 13, 2020.

9. Scheurer, D, Cardin T. The Role of NPs and PAs in Hospital Medicine Programs. The Hospitalist. July 14, 2017. www.the-hospitalist.org/hospitalist/article/142565/leadership-training/role-nps-and-pas-hospital-medicine-programs . Accessed December 13, 2017.

10. Metrics that Matter: Considering Clinical Complexity with Length of Stay More Accurately Indicates Efficiency. Care Logistics. August 22, 2016. www.carelogistics.com/blog/2016/8/22/metrics-that-matter-considering-clinical-complexity-with-length-of-stay-more-accurately-indicates-efficiency . Accessed May 20, 2019.

11. Rausch SL. The Pandemic Experience Through the Eyes of APPs. The Hospitalist. November 20, 2020. www.the-hospitalist.org/hospitalist/article/232240/coronavirus-updates/pandemic-experience-through-eyes-apps . Accessed December 19, 2020.