Operational Partners

Operational partners are system-level stakeholders who help ensure relevance of the review topic to the VA, contribute to the development of and approve final project scope and timeframe for completion, provide feedback on the draft report, and provide consultation on strategies for dissemination of the report to the field and relevant groups.

• Paul R. Conlin, M.D.
  Chair, Chiefs of Medicine Field Advisory Council
  Chief, Medical Service
  VA Boston Healthcare System
• Ajay Dhawan, M.D.
  Acting Chief Officer
  Specialty Care Services

Peer Reviewers

The Coordinating Center sought input from external peer reviewers to review the draft report and provide feedback on the objectives, scope, methods used, perception of bias, and omitted evidence (see Appendix E in Supplemental Materials for disposition of comments). Peer reviewers must disclose any relevant financial or non-financial conflicts of interest. Because of their unique clinical or content expertise, individuals with potential conflicts may be retained. The Coordinating Center works to balance, manage, or mitigate any potential nonfinancial conflicts of interest identified.

PURPOSE

The Evidence Synthesis Program (ESP) Coordinating Center is responding to a request from the VHA Office of Specialty Care Services (SCS) and Chiefs of Medicine Field Advisory Council, for an Evidence Brief on staffing models used in outpatient specialty care settings. Findings from this brief will be used to inform SCS and the Specialty Care Integrated Clinical Community in their support of field-based specialty care programs.

BACKGROUND

Staffing models are strategies to match the demand of tasks in an organization with the supply of staff, with the goals of minimizing staffing shortages, increasing the availability and use of staff expertise, and maximizing work outputs.¹ Tasks that are nonrepetitive in nature can be challenging to effectively staff; in healthcare settings in particular, nonrepetitive and complex tasks are prevalent and often reach across staff specializations. A common solution to addressing this challenge is the creation of healthcare teams.² Although teamwork in healthcare settings is associated with better staff performance (eg, improved processes or outcomes),³ research on and use of staffing models has generally been limited to inpatient and primary care settings. For example, a previous Cochrane review examined the relationships between inpatient nursing staffing levels/models and patient-important outcomes, but only found that staffing advanced or specialist nurses was not related to patient mortality.⁴

In primary care, the dominant staffing model is based in the patient-centered medical home (PCMH) care model. In PCMH, patients have a personal relationship with a primary care physician (PCP) who is responsible for all patient healthcare needs and for coordinating care across specialties.⁵ PCPs lead a team of, on average, 4 non-provider staff to collaboratively care for a panel of patients; non-provider staff can include nurse care managers, pharmacists, and nutritionists.⁶ In the VHA, the PCMH model was adapted into the patient-aligned care team (PACT),⁷ where a provider and 3 additional non-provider staff (1 registered nurse care manager, 1 licensed practical nurse or medical assistant, and 1 administrative clerk) are organized into a teamlet. As shown in Figure 1, the PACT is conceptually supported by 3 pillars: access, care management and coordination, and practice redesign; these elements are further supported by patient-centeredness, [system] improvement, and resources.⁸ The goals of each pillar are to increase appointment availability and non-appointment options (access), ensure that high-risk transitions are well-managed (care management and coordination), and to improve communication and work processes (practice redesign).

Figure 1

PACT Conceptual Model. Note. Figure adapted from Bidassie et al.

A meta-analysis of 19 PCMH interventions found that the model had small-to-moderate positive effects on the delivery of preventive care and a small positive effect on patient experience, with moderate strength of evidence.⁹ The review also found low strength of evidence for a small-to-moderate positive effect of PCMH on staff experience, and reduced emergency department visits. In the VHA, a nationwide observational study of 913 clinics, 5404 primary care staff, and 5.6 million Veterans found that clinics with better PACT implementation (ie, clinics that best adhere to the optimal PACT model) were associated with higher patient satisfaction, higher performance on clinical quality measures, lower staff burnout, and lower hospitalization and emergency department use.¹⁰ A rapid review of different team-based primary care structures found moderate strength of evidence for a nurse chronic care manager on improved patient outcomes, and low strength of evidence for nurse practitioner team co-managers on increased patient access, retrained medical assistants on increased screening rates, and variable team staff professions to match patient populations on higher quality of care.¹¹

Use of staffing models in outpatient specialty care settings may have similar benefits to productivity and patient-important outcomes by better matching staff to tasks appropriate for their skillsets, and by ensuring that staff are most efficiently utilized to avoid shortages. As in primary care, these models could make use of new or existing non-provider staff to complete different tasks or could be used to create new clinics to reorganize how all providers and staff collectively complete tasks to deliver patient care. No single staffing model is dominant in outpatient specialty care, but it is possible that similar models could be effective across specialties. In Table 1, we present possible outpatient specialty care staffing models or interventions, organized after the abovementioned Cochrane review⁴ on inpatient staffing models.

Table 1

Example Staffing Models/Interventions in Healthcare.

In summary, PCMH and PACT staffing models in primary care have been associated with increases in quality, access, patient and staff experience, and with fewer hospital admissions and emergency department visits. Compared with inpatient and primary care staffing models, however, the effects of different staffing models in outpatient specialty care are not well understood. The aim of this report is to summarize the literature on the relationships between outpatient specialty care staffing models, staff productivity, and patient-important outcomes, and describe how these models vary by program and contextual characteristics. In this Brief, we focus on supply-side changes to staffing (ie, new or different staff or clinics), rather than demand-side changes (ie, fewer or less complex patients).

PROTOCOL

A preregistered protocol for this review can be found on the PROSPERO international prospective register of systematic reviews (http://www.crd.york.ac.uk/PROSPERO/; registration number CRD42021285060).

KEY QUESTIONS

The following key questions were the focus of this review:

Key Question 1.

What staffing models for outpatient specialty care clinical support staff and clinicians are associated with increased staff productivity and improved patient outcomes?

Key Question 2.

How do staffing models for outpatient specialty care clinical support staff and clinicians vary by program or contextual characteristics (eg, program type, rural or urban setting, etc)?

ELIGIBILITY CRITERIA

The ESP included studies that met the following criteria:

DATA SOURCES AND SEARCHES

To identify articles relevant to the key questions, a research librarian searched Ovid MEDLINE, CINAHL, Cochrane Database of Systematic Reviews, and ClinicalTrials.gov as well as AHRQ and HSR&D databases through October 2021 using terms for outpatient specialty care programs (ie, immunology, oncology, cardiology), staffing and workload (see Appendix A in Supplemental Materials for complete search strategies). Additional citations were identified from hand-searching reference lists and consultation with content experts. We limited the search to published and indexed articles involving human subjects available in the English language. Study selection was based on the eligibility criteria described above. Models involving medical scribes were excluded, as a recent ESP systematic review¹² covered this intervention type in cardiology, orthopedic, and emergency departments. Titles, abstracts, and full-text articles were reviewed by 2 investigators. All disagreements were resolved by consensus or discussion with a third reviewer.

DATA ABSTRACTION AND ASSESSMENT

Effect information and population, intervention (or staffing model, if study involved no intervention), and comparator characteristics were abstracted from all included studies. The internal validity (risk of bias) of each included study was rated using the Cochrane Risk of Bias tools for randomized controlled trials¹³ and unrandomized studies,¹⁴ as well as the Joanna Briggs Institute (JBI) Checklist for cross-sectional and repeated cross-sectional studies.¹⁵ All data abstraction and internal validity ratings were first completed by 1 reviewer and then checked by another; disagreements were resolved by consensus or discussion with a third reviewer.

We graded the strength of the evidence (SOE) for each outcome based on the AHRQ Methods Guide for Comparative Effectiveness Reviews.¹⁶ This approach provides a rating of confidence in reported findings based on trial methodology (design, quality, and risk of bias), consistency (whether effects are in the same direction and have a consistent magnitude), and directness (whether assessed outcomes are clinically important to patients and providers). When information on precision of findings (eg, confidence intervals) is available, certainty of evidence is also evaluated. Because of substantial variation in outcomes across studies, we assessed strength of evidence for each intervention type/model rather than for each outcome type. For this review, high strength evidence consisted of multiple, large trials with low risk of bias (or good quality) and consistent and precise findings. Moderate strength evidence consisted of multiple trials with low or unclear risk of bias (mostly good or fair quality), and consistent and precise findings. Low strength evidence consisted of a single study or multiple small studies with unclear, some concerns for, or high risk of bias (or mostly fair or poor study quality) and inconsistent or imprecise findings. Insufficient evidence consisted of a single study or multiple small studies with unclear, some concerns for, or high risk of bias (or mostly fair or poor study quality) and inconsistent interventions/models and/or outcomes.

SYNTHESIS

We synthesized available evidence narratively by staffing model type (addition of new staff or redefined roles of existing staff, or establishment of a new clinic), describing patient and staffing characteristics and outcomes.

LITERATURE FLOW

The literature flow diagram (Figure 2) summarizes the results of the study selection process (full list of excluded studies available in Appendix B in Supplemental Materials).

Figure 2

Literature Flowchart. Abbreviation. CINAHL=Cumulative Index of Nursing and Allied Health.

LITERATURE OVERVIEW

Our search identified 5,068 potentially relevant articles. We included 8 studies in 10 publications,^17–26 which are summarized in Table 2 (see Appendix C in Supplemental Materials for full study details). Most studies^{17,20,22,24–26} examined the impact of new or additional mid-level staff or advanced practice providers (ie, nurses, advanced practice nurses [APNs], nurse practitioners [NPs], or physician assistants [PAs]) on productivity (ie, number of patients seen, ED visits or hospitalizations, or staff satisfaction) or patient outcomes (ie, patient satisfaction or symptom management). Two studies^21,23 examined the impact of new clinics developed to meet a specific clinical need (eg, symptom management among patients with breast cancer). Among the 5 studies reporting patient sample size, the median sample size was 199 (range: 55-15,381), with all but 2 studies including fewer than 200 patients. We identified 1 study in progress (see Appendix D in Supplemental Materials), which examined the use of a specially trained aid to support outpatient care for patients with congestive heart failure.

We excluded 51 studies after full-text screening for examining an ineligible intervention. Most of these studies evaluated interventions that were completely unrelated to staffing or work, but we did exclude a few intervention categories that were close to our topic but did not meet our inclusion criteria. These ineligible intervention categories included: telemedicine or telehealth interventions, descriptions of staffing interventions that did not include an evaluation, non-intervention measures of staffing, and non-staffing resource interventions (eg, changing patient registration procedures). We also excluded 32 studies for containing ineligible outcomes. Several of these studies did describe a staffing model or intervention but reported no eligible quantitative outcomes. The full list of excluded studies is available in Appendix B.

Of the studies that met inclusion criteria, most were cross-sectional^17,20,25 or repeated cross-sectional design.^21,24,26 Common limitations among the cross-sectional studies included lack of information on the patient population and lack of statistical analysis. Two cohort studies^22,23 were limited by lack of data provided on the samples, making it difficult to assess whether observed outcome changes were more likely due to the intervention or other sample characteristics.

No evidence was rated as high or moderate strength due to variability in staffing models and outcome measurement and study design limitations (eg, lack of comparison groups and/or statistical analysis). Our overall confidence in the evidence is low or insufficient due to limitations in study design (eg, lack of comparison group), other methodological limitations of the studies, and inconsistency in staffing models and outcome measurement. Additionally, lack of methodological detail made it difficult to evaluate reported associations between staffing models, productivity, and patient-important outcomes.

Table 2

Characteristics of Included Studies.

EFFECTIVENESS OF STAFFING MODELS FOR OUTPATIENT SPECIALTY CARE

New or Existing Staff

Three studies^24–26 that examined adding new staff to outpatient specialty clinics reported high patient, provider, and staff satisfaction and reductions in appointment wait times, inpatient admissions, and mortality (Table 3). In 1 study,²⁴ a new clinical team (1 physician, 2 NPs, 1 PA, supporting nursing, and medical staff) added to an outpatient neurology clinic to evaluate new neurology referrals and direct them either to primary care, a subspecialty clinic, or a general neurology physician, reduced monthly lead time (defined as time to third available appointment) from 299 days pre-implementation to 10 days post-implementation. Patient satisfaction was unchanged, with the same mean score pre- and post-implementation. In another study,²⁶ the addition of a physician extender (ie, NP or PA) to provide weekly visits, medication review, symptom management, and outreach reduced inpatient admissions, number of inpatient days, and mortality among outpatient chronic dialysis clinic patients. One study in an outpatient oncology setting²⁵ reported high levels of patient, provider, and staff satisfaction with the addition of an oncology pharmacist providing patient education, medication reconciliation, adverse drug event monitoring, and symptom management. Our confidence in these findings is low as they come from cross-sectional or repeated cross-sectional studies with limited information on the study samples and limited or no analysis of data.

Among studies examining existing staff (Table 4), 1 study¹⁷ assessed compliance to guideline-recommended heart failure measures among 167 outpatient cardiology practices based on the number of APNs or PAs on staff (0 vs >0 to <2 vs ≥2) and reported that practices with at least 2 APNs or PAs had greater compliance with 2 out of the 7 assessed measures compared to practices with less than 2 APNs or PAs; no significant differences were observed between practices in the other 5 measures. Another study²² reported reduced staffing violations after implementing a scheduling optimization intervention in an outpatient chemotherapy unit, and a final included study²⁰ linked nurse case manager activities to various quality measures, including improved continuity of care, patient self-management, and treatment adherence among complex hematologic cancer patients. As with studies of added staff, our confidence in these findings is low as they come from mostly cross-sectional or repeated cross-sectional studies with limited information on the study samples and limited or no analysis of data.

Table 3

Outcomes of Adding New Staff to an Existing Model of Care.

Table 4

Outcomes of Existing Staff on an Existing Model of Care.

New Clinics

Two studies^21,23 examined the effect of developing new clinics to meet specific needs within established outpatient clinics (Table 5).^21,23 In an outpatient oncology clinic, integrating a symptom management clinic led by oncology APNs into an existing oncology nurse triage system with the goal of enhancing coordination, communication, and patient education reduced oncology inpatient admissions and increased patient satisfaction scores. In another study,²³ a new headache infusion center staffed by RNs was able to reduce visit duration and cost by accommodating patients presenting to an outpatient neurology clinic meeting eligibility criteria for headache infusion with high patient satisfaction. However, we found these studies to provide insufficient evidence as they were limited by lack of information about the study and comparison populations, and lack of analysis.

Table 5

New Clinics Established Within an Existing Outpatient Setting.

STAFFING MODEL CONTEXTUAL CHARACTERISTICS

We did not identify any studies that compared different staffing models head-to-head, so we cannot make any direct conclusions about the relationship between staffing models and contextual characteristics (eg, specialty, rurality, adding staff vs clinics, etc). Studies did vary by specialty and rurality, but indirect comparisons of these characteristics across studies also could not be made for most outcomes, as they were not consistent between studies. However, patient, provider, and staff satisfaction were high or positive across studies of different specialties. Patient satisfaction was frequently measured (4 studies),^21,23–25 and was consistently high. Provider and staff satisfaction, measured in 2 studies,^22,25 was high in 1 study, and positive but not quantitatively reported in the other.

LIMITATIONS

The studies found in this review were mostly of fair or poor quality, with only 1 employing a design that utilized a comparison group.²³ Without a comparison group, it was not possible to determine if the evaluated staffing models perform better than usual care. Further, only 1¹⁷ of the 8 included studies^17,20–26 employed any type of statistical analysis. Such analyses may or may not have shown intervention effects, but without them, the strength of evidence is much lower. Studies without statistical analyses presented interventions and findings in a case study fashion that omitted many important details (eg, patient or staff characteristics), as noted in the study quality ratings (see Appendix C in the Supplemental Materials).

FUTURE RESEARCH

There is need for research on outpatient specialty care staffing in the VHA to examine the addition of a specific type of staff member or clinic across specialties. At present, the existing literature lacks a uniform model of outpatient specialty care staffing. It is unclear if this is due to different specialty needs or a paucity of research on staffing models. Research that focuses on analyzing the effects of advanced practice providers/clinical support staff or triage clinics on productivity and patient outcomes in specialty care could help the VHA optimize its staffing models. Stakeholders in VA should first define a conceptual model of specialty care delivery, like PACT in primary care, and then map ideal staffing and clinics to that model; researchers can then follow this conceptual model as a map to evaluate different staffing or clinic models to determine the best way to deliver specialty care. Without an overarching conceptual model, it would be difficult to direct future work in this area.

In addition, although case studies can be useful for management and practitioners, their findings are difficult to generalize. More rigorous staffing research, including studies with randomized interventional designs (such as the randomized evaluation model used by the VA’s Partnered Evidence-based Policy Resource Center), relevant comparators, and guidance by a specialty care conceptual model, may aid the VHA in identifying effective outpatient specialty care staffing models. In addition, the COVID-19 pandemic, and the ensuing healthcare staffing shortages and shift towards telemedicine, has created new challenges for specialty care staffing. Future research should examine how changes to staffing models can respond to short-term staffing shortages or the need to maintain both in-person and virtual clinics. Even as the pandemic recedes, planning for responsive staffing models should be ongoing to ensure preparedness for future care disruptions.

Finally, many VA facilities and clinics, like other major healthcare systems with teaching hospitals, utilize healthcare trainees (including physician residents and fellows) to assist in providing patient care while receiving training. Research should examine how these trainees can augment existing specialty care staffing levels, while still maintaining quality of care and maximizing trainee education.

CONCLUSIONS

Little evidence is available on staffing models in the outpatient specialty care context. Existing models are generally ad hoc in nature, and research to date has lacked an overarching conceptual model analogous to those used in primary care settings and has been hampered by methodological inconsistency and other limitations. In the VA context, development and implementation of specialty care staffing models that improve productivity and patient outcomes would be facilitated through use of 1) an overarching conceptual framework that identifies key staffing model elements and outcomes across specialty care settings, and 2) more rigorous study designs and analysis methods that would increase the informativeness of evidence on specialty care staffing models

APPENDIX A. SEARCH STRATEGY

APPENDIX B. EXCLUDED STUDIES

Exclude reasons: 1=Ineligible population, 2=Ineligible intervention, 3=Ineligible comparator, 4=Ineligible outcome, 5=Ineligible timing, 6=Ineligible study design, 7=Ineligible publication type 8=Outdated or ineligible systematic review, 9=Non-English language, 10=Unable to obtain full text

APPENDIX C. EVIDENCE TABLES

APPENDIX D. RESEARCH IN PROGRESS

APPENDIX E. PEER REVIEWER DISPOSITION

REFERENCES

1.

Albert NM, Fonarow GC, Yancy CW, et al. Outpatient cardiology practices with advanced practice nurses and physician assistants provide similar delivery of recommended therapies (findings from IMPROVE HF). American Journal of Cardiology. 2010;105(12):1773–1779. [PubMed: 20538129]

2.

Garnett D, Hardy L, Fitzgerald E, Fisher T, Graham L, Overcash J. Nurse Case Manager: Measurement of Care Coordination Activities and Quality and Resource Use Outcomes When Caring for the Complex Patient With Hematologic Cancer. Clinical Journal of Oncology Nursing. 2020;24(1):65–74. [PubMed: 31961839]

3.

Graze L, Brady-Copertino C, Varner A, Stiver WS. The development of a nursing assessment and symptom management clinic. Clinical Journal of Oncology Nursing. 2014;18 Suppl:12–16. [PubMed: 25252985]

4.

Huang YL, Bryce AH, Culbertson T, et al. Alternative Outpatient Chemotherapy Scheduling Method to Improve Patient Service Quality and Nurse Satisfaction. Journal of oncology practice/American Society of Clinical Oncology. 2018;14(2):e82–e91. [PMC free article: PMC6550030] [PubMed: 29272201]

5.

Reese O, Wright C, Sandoval A, et al. “Establishment of an acute headache infusion clinic as an alternative for emergency department care”. Journal of the Neurological Sciences. 2021;423:117384. [PubMed: 33714085]

6.

Ross SC. An option for improving access to outpatient general neurology. Neurology Clinical Practice. 2014;4(5):435–440. [PMC free article: PMC5765689] [PubMed: 29443242]

7.

Ruder AD, Smith DL, Madsen MT, Kass FH, 3rd. Is there a benefit to having a clinical oncology pharmacist on staff at a community oncology clinic? Journal of Oncology Pharmacy Practice. 2011;17(4):425–432. [PubMed: 21248174]

8.

Smith GO, Jr. The role of physician assistants in improving renal care. Nephrology News & Issues. 2004;18(5):51–56. [PubMed: 15125166]

9.

Fonarow GC, Albert NM, Curtis AB, et al. Improving evidence-based care for heart failure in outpatient cardiology practices: primary results of the Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF). In. Circulation. Vol 1222010:585–596. [PubMed: 20660805]

10.

Fessele K, Yendro S, Mallory G. Setting the bar: developing quality measures and education programs to define evidence-based, patient-centered, high-quality care. Clinical Journal of Oncology Nursing. 2014;18. [PubMed: 25252984]