Stages of change: Strategies to promote use of a Pediatric Early Warning System in resource-limited pediatric oncology centers

BACKGROUND

Pediatric Early Warning Systems (PEWS) assist early detection of clinical deterioration in hospitalized children with cancer. Relevant to successful PEWS implementation, the "stages of change" model characterizes stakeholder support for PEWS based on willingness and effort to adopt the new practice.

METHODS

At five resource-limited pediatric oncology centers in Latin America, semi-structured interviews were conducted with 71 hospital staff involved in PEWS implementation. Purposive sampling was used to select centers requiring variable time to complete PEWS implementation, with low-barrier centers (3-4 months) and high-barrier centers (10-11 months). Interviews were conducted in Spanish, professionally transcribed, and translated into English. Thematic content analysis explored "stage of change" with constant comparative analysis across stakeholder types and study sites.

RESULTS

Participants identified six interventions (training, incentives, participation, evidence, persuasion, and modeling) and two policies (environmental planning and mandates) as effective strategies used by implementation leaders to promote stakeholder progression through stages of change. Key approaches involved presentation of evidence demonstrating PEWS effectiveness, persuasion and incentives addressing specific stakeholder interests, enthusiastic individuals serving as models for others, and policies enforced by hospital directors facilitating habitual PEWS use. Effective engagement targeted hospital directors during early implementation phases to provide programmatic legitimacy for clinical staff.

CONCLUSION

This study identifies strategies to promote adoption and maintained use of PEWS, highlighting the importance of tailoring implementation strategies to the motivations of each stakeholder type. These findings can guide efforts to implement PEWS and other evidence-based practices that improve childhood cancer outcomes in resource-limited hospitals.

The COVID-19 Pandemic's impact on sustainability and expansion of a Pediatric Early Warning System in resource-limited hospitals

BACKGROUND

The COVID-19 pandemic impacted healthcare delivery worldwide, including pediatric cancer care, with a disproportionate effect in resource-limited settings. This study evaluates its impact on existing quality improvement (QI) programs.

METHODS

We conducted 71 semi-structured interviews of key stakeholders at five resource-limited pediatric oncology centers participating in a collaborative to implement Pediatric Early Warning System (PEWS). Interviews were conducted virtually using a structured interview guide, recorded, transcribed, and translated into English. Two coders developed a codebook of a priori and inductive codes and independently coded all transcripts, achieving a kappa of 0.8-0.9. Thematic analysis explored the impact of the pandemic on PEWS.

RESULTS

All hospitals reported limitations in material resources, reduction in staffing, and impacts on patient care due to the pandemic. However, the impact on PEWS varied across centers. Identified factors that promoted or limited ongoing PEWS use included the availability of material resources needed for PEWS, staff turnover, PEWS training for staff, and the willingness of staff and hospital leaders to prioritize PEWS. Consequently, some hospitals were able to sustain PEWS; others halted or reduced PEWS use to prioritize other work. Similarly, the pandemic delayed plans at all hospitals to expand PEWS to other units. Several participants were hopeful for future expansion of PEWS post-pandemic.

CONCLUSION

The COVID-19 pandemic created challenges for sustainability and scale of PEWS, an ongoing QI program, in these resource-limited pediatric oncology centers. Several factors mitigated these challenges and promoted ongoing PEWS use. These results can guide strategies to sustain effective QI interventions during future health crises.

Impact of hospital characteristics on implementation of a Pediatric Early Warning System in resource-limited cancer hospitals

Background

Pediatric Early Warning Systems (PEWS) aid in identification of deterioration in hospitalized children with cancer but are underutilized in resource-limited settings. Proyecto EVAT is a multicenter quality improvement (QI) collaborative in Latin America to implement PEWS. This study investigates the relationship between hospital characteristics and time required for PEWS implementation.

Methods

This convergent mixed-methods study included 23 Proyecto EVAT childhood cancer centers; 5 hospitals representing quick and slow implementers were selected for qualitative analysis. Semi-structured interviews were conducted with 71 stakeholders involved in PEWS implementation. Interviews were recorded, transcribed and translated to English, then coded using a priori and novel codes. Thematic content analysis explored the impact of hospital characteristics and QI experience on time required for PEWS implementation and was supplemented by quantitative analysis exploring the relationship between hospital characteristics and implementation time.

Results

In both quantitative and qualitative analysis, material and human resources to support PEWS significantly impacted time to implementation. Lack of resources produced various obstacles that extended time necessary for centers to achieve successful implementation. Hospital characteristics, such as funding structure and type, influenced PEWS implementation time by determining their resource-availability. Prior hospital or implementation leader experience with QI, however, helped facilitate implementation by assisting implementers predict and overcome resource-related challenges.

Conclusions

Hospital characteristics impact time required to implement PEWS in resource-limited childhood cancer centers; however, prior QI experience helps anticipate and adapt to resource challenges and more quickly implement PEWS. QI training should be a component of strategies to scale-up use of evidence-based interventions like PEWS in resource-limited settings.

Multilevel impacts of a pediatric early warning system in resource-limited pediatric oncology hospitals

Background

Pediatric Early Warning Systems (PEWS) reduce clinical deterioration, improve interdisciplinary communication, and provide cost savings; however, little is known about how these impacts are achieved or related. This study evaluates the multi-level impacts of PEWS in resource-limited pediatric oncology centers.

Methods

We conducted 71 semi-structured interviews including physicians (45%), nurses (45%), and administrators (10%) from 5 resource-limited pediatric oncology centers in 4 Latin American countries. Interviews were conducted in Spanish, transcribed, and translated into English. A code book was developed using a priori and inductively derived codes. Transcripts were independently coded by 2 coders, achieving a kappa of 0.8-0.9. Thematic content analysis explored perceived impacts of PEWS at the level of the patient, clinician, healthcare team, and institution.

Results

PEWS improved the quality of attention for patients, reducing morbidity and mortality. Clinicians felt more knowledgeable, confident, and empowered providing patient care, resulting in greater job satisfaction. PEWS affected team dynamics by improving interdisciplinary (ward and intensive care unit) and interprofessional (physicians and nurses) relationships and communication. This ultimately led to institutional culture change with emphasis on patient safety, collaboration with other centers, and receipt of institutional awards. Together, these impacts led to hospital-wide support of ongoing PEWS use.

Conclusions

In resource-limited hospitals, PEWS use results in multi-level positive impacts on patients, clinicians, teams, and institutions, creating a feedback loop that further supports ongoing PEWS use. These findings can guide advocacy for PEWS to various stakeholders, improve PEWS effectiveness, and inform assessment of other interventions to improve childhood cancer outcomes.

Assessment of Barriers and Enablers to Implementation of a Pediatric Early Warning System in Resource-Limited Settings

IMPORTANCE

Pediatric early warning systems (PEWS) aid with early identification of clinical deterioration and improve outcomes in children with cancer hospitalized in resource-limited settings; however, there may be barriers to implementation.

OBJECTIVE

To evaluate stakeholder-reported barriers and enablers to PEWS implementation in resource-limited hospitals.

DESIGN, SETTING, AND PARTICIPANTS

In this qualitative study, semistructured stakeholder interviews were conducted at 5 resource-limited pediatric oncology centers in 4 countries in Latin America. Hospitals participating in a multicenter collaborative to implement PEWS were purposefully sampled based on time required for implementation (fast vs slow), and stakeholders interviewed included physicians, nurses, and administrators, involved in PEWS implementation. An interview guide was developed using the Consolidated Framework for Implementation Research (CFIR). Interviews were conducted virtually in Spanish, audiorecorded, and professionally transcribed and translated into English. A codebook was developed a priori using the CFIR and supplemented with codes inductively derived from transcript review. Two coders independently analyzed all transcripts, achieving a κ of 0.8 to 0.9. The study was conducted from June 1 to August 31, 2020.

MAIN OUTCOMES AND MEASURES

Thematic analysis was conducted based on CFIR domains (inner setting, characteristics of individuals, outer setting, intervention characteristics, and implementation process) to identify barriers and enablers to PEWS implementation.

RESULTS

Seventy-one staff involved in PEWS implementation were interviewed, including 32 physicians (45%), 32 nurses (45%), and 7 administrators (10%). Of these, 50 were women (70%). Components of the 5 CFIR domains were mentioned by participants as barriers and enablers to PEWS implementation at both fast- and slow-implementing centers. Participants emphasized barriers at the level of the clinical staff, hospital, external factors, and PEWS intervention. These barriers included staff resistance to change, inadequate resources, components of health systems, and the perceived origin and complexity of PEWS. At all centers, most barriers were successfully converted to enablers during the implementation process through targeted strategies, such as early stakeholder engagement and adaptation, including adapting PEWS to better fit the local context and changing the hospital setting to support ongoing use of PEWS.

CONCLUSIONS AND RELEVANCE

To date, this is the first multicenter, multinational study describing barriers and enablers to PEWS implementation in resource-limited settings. Findings suggest that many barriers are not immutable and can be converted to enablers during the implementation process. This work can serve as a guide for clinicians looking to implement evidence-based interventions to reduce global disparities in patient outcomes.

Challenges to sustainability of pediatric early warning systems (PEWS) in low-resource hospitals in Latin America

Background

Sustainability, or continued use of evidence-based interventions for long-term patient benefit, is the least studied aspect of implementation science. In this study, we evaluate sustainability of a Pediatric Early Warning System (PEWS), an evidence-based intervention to improve early identification of clinical deterioration in hospitalized children, in low-resource settings using the Clinical Capacity for Sustainability Framework (CCS).

Methods

We conducted a secondary analysis of a qualitative study to identify barriers and enablers to PEWS implementation. Semi-structured interviews with PEWS implementation leaders and hospital directors at 5 Latin American pediatric oncology centers sustaining PEWS were conducted virtually in Spanish from June to August 2020. Interviews were recorded, professionally transcribed, and translated into English. Exploratory thematic content analysis yielded staff perceptions on PEWS sustainability. Coded segments were analyzed to identify participant perception about the current state and importance of sustaining PEWS, as well as sustainability successes and challenges. Identified sustainability determinants were mapped to the CCS to evaluate its applicability.

Results

We interviewed 71 staff including physicians (45%), nurses (45%), and administrators (10%). Participants emphasized the importance of sustaining PEWS for continued patient benefits. Identified sustainability determinants included supportive leadership encouraging ongoing interest in PEWS, beneficial patient outcomes enhancing perceived value of PEWS, integrating PEWS into the routine of patient care, ongoing staff turnover creating training challenges, adequate material resources to promote PEWS use, and the COVID-19 pandemic. While most identified factors mapped to the CCS, COVID-19 emerged as an additional external sustainability challenge. Together, these challenges resulted in multiple impacts on PEWS sustainment, ranging from a small reduction in PEWS quality to complete disruption of PEWS use and subsequent loss of benefits to patients. Participants described several innovative strategies to address identified challenges and promote PEWS sustainability.

Conclusion

This study describes clinician perspectives on sustainable implementation of evidence-based interventions in low-resource settings, including sustainability determinants and potential sustainability strategies. Identified factors mapped well to the CCS, however, external factors, such as the COVID pandemic, may additionally impact sustainability. This work highlights an urgent need for theoretically-driven, empirically-informed strategies to support sustainable implementation of evidence-based interventions in settings of all resource-levels.

Radiofrequency minimally invasive endovascular treatment of lower limbs varicose veins: clinical experience and literature review

AIM

Varicose veins of the legs are a common condition affecting 10-15% of men and 20-25% of women in the western world. This high prevalence is responsible of high medical and social costs. Most primary varices are associated with greater saphenous vein (GSV) incompetence. A new method, radiofrequency (RF) endovenous obliteration (VNUS-Closure'' procedure), recently has been described as a less invasive and cost-saving alternative to stripping for the treatment of refluxing GSV.

METHODS

Twenty-four patients with varicose veins underwent endovenous obliteration of the above knee GSV by VNUS Closure'' procedure. The vein diameters were from 5 to 10 mm. The RF catheter was inserted via percutaneous puncture or through a small skin incision. All operations were performed in local, tumescent anesthesia, under ultrasound guidance. All patients were discharged 2 h after operation. Clinical and ultrasound follow-up was performed at 1 week, and at 1, 6, 12, 24 months.

RESULTS

The complete or partial occlusion of the treated segment of the GSV has been achieved in 23 cases. In only one patient persisting patency of the GSV was immediately detected after the procedure. That was successfully treated by ultrasound guided foam sclerotherapy. All patients could resume all normal activities within 3-5 days. Every patient had reduction of varicosities, leg pain, fatigue and oedema. Adverse sequelae were minimal: 2 patients had transient thigh paresthesias. We didn't report deep venous thrombosis or pulmonary embolism (mean follow-up 26.7 months, range 15-33 months).

CONCLUSION

A literature review and the authors'experience reveal that, in absence of significant complications, such as deep vein thrombosis and pulmonary embolism, there are significant advantages in the RF endovascular obliteration of the GSV. In effect, the Closure'' procedure, in selected patients, offers reduced postoperative pain, shorter sick leaves, faster return to normal activities compared with vein stripping, and it appears to be cost-saving for society. The mid-term (36 months) recurrence rates after RF obliteration seem to be similar to the results of the conventional surgical management.

Contrasting interspecies efficacy and toxicology of 1,2-diethyl-3-hydroxypyridin-4-one, CP94, relates to differing metabolism of the iron chelating site

In order to define a predictive animal model for the effects of hydroxypyridinone (HPO) iron chelators in humans, we have compared the 28 d oral efficacy and toxicology of the HPO, 1,2-diethyl-3-hydroxypyridin-4-one (CP94) in rats and guinea-pigs and related the results to the contrasting metabolism of this compound in the two species. CP94 was highly effective at mobilizing liver iron in rats but showed toxicity at higher doses, whereas in the guinea-pig the compound lacked toxicity but was ineffective at mobilizing liver iron. These differences can be explained by the contrasting metabolism of the drug between the two species. In rats, at the top dose of 300 mg/kg intragastrically, all animals died before the end of the study, with no deaths or weight loss at lower doses. At 100 mg/kg, rat liver non-haem iron concentrations were reduced by 53% and 44% in females and males respectively (P < 0.001). At this dose, adrenal medullary cell vacuolation, increased mammary secretory activity, vacuolation of corpora luteal cells and single cell hepatocyte necrosis were seen. There were no reductions in the white cell count. At 50 mg/kg rat liver non-haem iron concentrations were decreased by 50% and 34% in females and males respectively (P < 0.02). In female rats this was associated with increased mammary secretory activity. In iron-overloaded rats given 100 mg/kg by gavage for 28 d, liver non-haem iron concentration was reduced by 39% (P < 0.01) and serum ferritin by 71% (P < 0.001). Ovarian and mammary changes were not influenced by iron loading. In guinea-pigs, CP94 was evaluated at 50 mg/kg, 100 mg/kg or 200 mg/kg by oral insufflation for 28 d. No reduction in liver iron was seen and no systematic dose related histological, biochemical or haematological effects were observed. Whereas in guinea-pigs 99% of urinary recovery following an oral dose of CP94 (100 mg/kg) was as the inactive glucuronide metabolite, in the rat only 23% of the dose was excreted in the urine as the glucuronide with remainder as the free drug or an iron binding metabolite. The lack of both efficacy and toxicity in the guinea-pig may therefore be explained by the rapid inactivation of CP94 by glucuronidation. This metabolism of CP94 in the guinea-pig is closer to humans than the rat, suggesting that both the efficacy and toxicity of this compound in humans may also be limited by glucuronidation.

Abrogation of H-2-restricted CTL responses and efficient recognition of HLA-A3 molecules in DBA/2 HLA/A24 responder mice

DBA/2 (H-2d) HLA-B7 x human beta 2-microglobulin transgenic and DBA/2 nontransgenic mice stimulated by DBA/2 HLA-A3 x human beta 2-microglobulin transgenic mouse spleen cells developed potent H-2Kd restricted cytolytic responses with recognition of a peptide from the second domain of the HLA-A3 H chain. These H-2Kd-restricted responses obliterated, as a rule, cytolytic responses with direct recognition of the HLA-A3 molecules, even in HLA-B7 transgenic mice. These immunodominant H-2Kd-restricted responses could be abrogated in DBA/2 HLA-A24 mice because of cross-tolerance, the HLA-A3 derived-H-2Kd presented peptide being shared by several (including A24) HLA class I H chain allelic variants. Under such experimental circumstances, strong CTL responses with exclusive direct recognition of HLA-A3 molecules constantly developed. Further analysis of these responses in six DBA/2 HLA-A24 responder mice indicated that a large fraction of the mouse V beta and V alpha genes could be used to mount such CTL responses. Thus, by combining classical HLA class I transgenesis and selective destruction of H-2K and H-2D genes, it should be possible to derive useful strains of mice for the study of HLA class I-restricted CTL responses.

Abrogation of H-2-restricted CTL responses and efficient recognition of HLA-A3 molecules in DBA/2 HLA/A24 responder mice

DBA/2 (H-2d) HLA-B7 x human beta 2-microglobulin transgenic and DBA/2 nontransgenic mice stimulated by DBA/2 HLA-A3 x human beta 2-microglobulin transgenic mouse spleen cells developed potent H-2Kd restricted cytolytic responses with recognition of a peptide from the second domain of the HLA-A3 H chain. These H-2Kd-restricted responses obliterated, as a rule, cytolytic responses with direct recognition of the HLA-A3 molecules, even in HLA-B7 transgenic mice. These immunodominant H-2Kd-restricted responses could be abrogated in DBA/2 HLA-A24 mice because of cross-tolerance, the HLA-A3 derived-H-2Kd presented peptide being shared by several (including A24) HLA class I H chain allelic variants. Under such experimental circumstances, strong CTL responses with exclusive direct recognition of HLA-A3 molecules constantly developed. Further analysis of these responses in six DBA/2 HLA-A24 responder mice indicated that a large fraction of the mouse V beta and V alpha genes could be used to mount such CTL responses. Thus, by combining classical HLA class I transgenesis and selective destruction of H-2K and H-2D genes, it should be possible to derive useful strains of mice for the study of HLA class I-restricted CTL responses.

T cell receptor selection by and recognition of two class I major histocompatibility complex-restricted antigenic peptides that differ at a single position

Peptides derived from HLA-Cw3 and HLA-A24 within region 170-179 differ by a single substitution, at position 173, and are both presented by the class I major histocompatibility complex molecule H-2Kd for recognition by murine cytolytic T lymphocytes (CTLs). As a first approach to understand the way T cell receptors (TCRs) intact with the HLA peptides, we have analyzed the TCR selection by, and recognition of, the two HLA antigenic sites. First, we have compared the TCR repertoires selected by HLA-Cw3 and HLA-A24, not only by sequencing the TCRs carried by CTL clones isolated and grown in vitro, but also by analyzing the TCRs expressed in vivo by peritoneal exudate lymphocytes from immune animals. Second, we have compared the TCR crossrecognition of HLA-A24 by CTLs selected by HLA-Cw3 with that of HLA-Cw3 by CTLs selected by HLA-A24. The combined analysis of TCR selection by and recognition of these two related HLA antigenic sites provides evidence that the TCR beta junctional regions interact with the amino-terminal part of the HLA peptides.

H-2-restricted cytolytic T lymphocytes specific for HLA display T cell receptors of limited diversity

We previously showed that H-2Kd-restricted cytotoxic T lymphocyte (CTL) clones specific for a single nonapeptide derived from the Plasmodium berghei circumsporozoite (PbCS) protein displayed T cell receptors (TCRs) of highly diverse primary structure. We have now analyzed the TCR repertoire of CTLs that recognize a peptide derived from the human class I major histocompatibility complex (MHC) molecule HLA-Cw3 in association with the same murine class I MHC molecule H-2Kd. We first sequenced the TCR alpha and beta genes of the CTL clone Cw3/1.1 and, based on this genomic analysis, the TCR alpha and beta cDNA junctional regions of 23 independent H-2Kd-restricted CTL clones specific for HLA-Cw3. The results show that the TCR chains display very limited heterogeneity, both in terms of V alpha, J alpha, V beta, and J beta segments, and in terms of length and sequence of the CDR3 alpha and beta loops. The TCR repertoire used in vivo was then analyzed by harvesting CTL populations from the peritoneal cavity of immune mice. The peritoneal exudate lymphocytes (PELs) displayed HLA-Cw3-specific cytolytic activity in the absence of any stimulation in vitro. Remarkably, most of these freshly isolated PELs expressed TCRs that shared the same structural features as those from HLA-Cw3-reactive CTL clones. Thus, our results show that a peptide from HLA-Cw3 presented by H-2Kd selects CTLs that bear TCRs of very limited diversity in vivo. When taken together with the high diversity of the TCRs specific for the PbCS peptide, these findings suggest that natural tolerance to self peptides presented by class I MHC molecules may substantially reduce the size of the TCR repertoire of CTLs specific for antigenic peptides homologous to self.

Analysis of the HLA-Cw3-specific cytotoxic T lymphocyte response of HLA-B7 X human beta 2m double transgenic mice

The cytolytic responses of either normal (non transgenic), HLA-B7 (single transgenic) or HLA-B7 x human beta 2 microglobulin (double transgenic) DBA/2 mice induced by transfected HLA-Cw3 P815 (H-2d) mouse mastocytoma cells were compared, to evaluate whether the expression of an HLA class I molecule in responder mice would favor the emergence of HLA-specific, H-2-unrestricted CTL. Only 8 of 300 HLA-Cw3-specific CTL clones tested could selectively lyse HLA-Cw3-transfected cells in an H-2-unrestricted manner, all having been isolated after hyperimmunization of double transgenic mice. These clones also lysed HLA-Cw3+ human cells. Unexpectedly, the lysis of the human but not that of the murine HLA-Cw3 cells was inhibited by Ly-2,3-specific mAb. Despite significant expression of HLA-B7 class I molecules on transgenic lymphoid cells, including thymic cells, limiting dilution analysis and comparative study of TCR-alpha and -beta gene rearrangements of the eight isolated clones (which suggested that they all derived from the same CTL precursor) indicated that the frequency of HLA-Cw3-specific H-2 unrestricted cytotoxic T lymphocytes remained low (even in HLA-B7 x human beta 2-microglobulin double transgenic mice). This suggests that coexpression of HLA class I H and L chain in transgenic mice is not the only requirement for significant positive selection of HLA class I-restricted cytotoxic mouse T lymphocytes.

Analysis of the HLA-Cw3-specific cytotoxic T lymphocyte response of HLA-B7 X human beta 2m double transgenic mice

The cytolytic responses of either normal (non transgenic), HLA-B7 (single transgenic) or HLA-B7 x human beta 2 microglobulin (double transgenic) DBA/2 mice induced by transfected HLA-Cw3 P815 (H-2d) mouse mastocytoma cells were compared, to evaluate whether the expression of an HLA class I molecule in responder mice would favor the emergence of HLA-specific, H-2-unrestricted CTL. Only 8 of 300 HLA-Cw3-specific CTL clones tested could selectively lyse HLA-Cw3-transfected cells in an H-2-unrestricted manner, all having been isolated after hyperimmunization of double transgenic mice. These clones also lysed HLA-Cw3+ human cells. Unexpectedly, the lysis of the human but not that of the murine HLA-Cw3 cells was inhibited by Ly-2,3-specific mAb. Despite significant expression of HLA-B7 class I molecules on transgenic lymphoid cells, including thymic cells, limiting dilution analysis and comparative study of TCR-alpha and -beta gene rearrangements of the eight isolated clones (which suggested that they all derived from the same CTL precursor) indicated that the frequency of HLA-Cw3-specific H-2 unrestricted cytotoxic T lymphocytes remained low (even in HLA-B7 x human beta 2-microglobulin double transgenic mice). This suggests that coexpression of HLA class I H and L chain in transgenic mice is not the only requirement for significant positive selection of HLA class I-restricted cytotoxic mouse T lymphocytes.

Cytotoxic T lymphocyte recognition of secreted HLA class I molecules

The cytolytic responses of DBA/2 mice against syngeneic transfected P815 mastocytoma cells expressing either membrane-associated (HLA-Cw3) or -secreted hybrid (HLA-Cw3 x H-2 Q10b) molecules were compared. In spite of the absence of serologically detectable hybrid molecules on their plasma membrane, cells secreting these molecules elicited a CTL response similar to that of cells expressing the membrane associated HLA-Cw3 molecules, in terms of both MHC-restriction and peptide specificity. Together with the observation that syngeneic mice were capable of rejecting the injected secreting cells, these results imply that secreted HLA class I molecules can function as minor histocompatibility Ag and suggest that processing of both the membrane-bound and the -secreted forms of a protein may follow common or overlapping pathways.

Cytotoxic T lymphocyte recognition of secreted HLA class I molecules

The cytolytic responses of DBA/2 mice against syngeneic transfected P815 mastocytoma cells expressing either membrane-associated (HLA-Cw3) or -secreted hybrid (HLA-Cw3 x H-2 Q10b) molecules were compared. In spite of the absence of serologically detectable hybrid molecules on their plasma membrane, cells secreting these molecules elicited a CTL response similar to that of cells expressing the membrane associated HLA-Cw3 molecules, in terms of both MHC-restriction and peptide specificity. Together with the observation that syngeneic mice were capable of rejecting the injected secreting cells, these results imply that secreted HLA class I molecules can function as minor histocompatibility Ag and suggest that processing of both the membrane-bound and the -secreted forms of a protein may follow common or overlapping pathways.