Category: Statistics

BMC Health Serv Res. 2026 Jul 18. doi: 10.1186/s12913-026-15127-4. Online ahead of print.

ABSTRACT

INTRODUCTION: Kenya introduced the Managed Equipment Services (MES) model in 2015 as a seven-year national contract to address significant gaps in access to hemodialysis, which had previously been available in fewer than ten public hospitals. The MES aimed to expand hemodialysis capacity by centralizing procurement, standardizing equipment, and ensuring maintenance across all 47 counties. This study provides the first national evaluation of hemodialysis performance under the MES model.

METHODS: We conducted a multi-year retrospective analysis of 54 public hospitals equipped through the MES, six years after implementation, using data from the MES contract, the MES final survey, NHIF reimbursement schedules, and the 2019 Population Census. Hemodialysis system performance was assessed across four domains, including capacity, utilization, reporting accuracy, and cost efficiency, using descriptive statistics, time series trends, and discrepancy analysis.

RESULTS: A total of 54 hemodialysis centers equipped with 305 machines were established across all 47 counties, resulting in national densities of 1.14 centers and 6.41 machines per million people. Machine density varied more than 30-fold across counties. Between June 2015 and April 2021, 392,497 hemodialysis sessions were conducted, with monthly utilization increasing by 376%, from 2,100 sessions in June 2015 to 10,173 in April 2021. Time-series regression showed a significant upward trend (+75.5 sessions per month, R² = 0.721), with significant effects of year (F = 258.42, p < 0.001) and calendar month (F = 5.00, p < 0.001), and two periods of acceleration in June 2016 and January 2020. Facility throughput ranged from 217 to 79,202 sessions, and 63% of hospitals operated at less than 50% of their machine capacity. Machine-logged session counts exceeded register-based counts by more than 30% in 79% of hospitals, with a mean absolute percentage error of 180.41%. County-level utilization ranged from 89 to 4,363 sessions per million people, a 49-fold difference. The MES program reached breakeven in month 70, with a projected return on investment of 19.76% by month 84.

CONCLUSION: The MES significantly increased Kenya’s public hemodialysis capacity and supported rapid nationwide growth in service use. However, disparities in geography, facilities, and documentation persisted, with many hospitals operating below capacity and inconsistent reporting. Although the program broke even financially and had a positive return on investment, facility performance varied widely. These results show that expanding infrastructure alone was not enough to achieve equity or efficient hemodialysis service delivery, highlighting the need to improve staffing, data quality, and system coordination to make the most of future renal care investments.

PMID:42471648 | DOI:10.1186/s12913-026-15127-4

BMC Pulm Med. 2026 Jul 18. doi: 10.1186/s12890-026-04497-4. Online ahead of print.

ABSTRACT

BACKGROUND: Plasma transfusion is used in patients with severe pneumonia who develop coagulopathy or bleeding, but recipients are often sicker than non-recipients. We evaluated the association between plasma transfusion and in-hospital mortality after adjustment for available measured confounders, while recognizing that unmeasured confounding from transfusion indications and illness severity may persist.

METHODS: This single-center retrospective cohort included 156 adults with severe pneumonia admitted between January 2021 and May 2024; 62 received plasma and 94 did not. The matched mortality comparison using propensity score matching (PSM) and McNemar analysis was prespecified as the primary analysis. Supportive and exploratory analyses included Cox regression, Kaplan-Meier analysis, IPTW, overlap weighting, Fine-Gray modeling, E-value analysis, and exploratory coagulation-enhanced, logistic, microbiological/coagulation, and time-dependent exposure sensitivity analyses.

RESULTS: Before matching, mortality was higher among plasma recipients than controls (64.5% vs. 28.0%; crude odds ratio [OR] 4.69, 95% confidence interval [CI] 2.35-9.34; p < 0.001), with substantial baseline imbalance. After matching, mortality remained numerically higher in the plasma group (50.0% vs. 36.1%) but was not statistically significant (matched OR 1.62, 95% CI 0.67-3.92; McNemar p = 0.383). Cox models yielded attenuated HRs after adjustment, but the fully adjusted model violated the global proportional hazards assumption; therefore, Cox estimates were interpreted as time-weighted associations over follow-up rather than constant effects. Exploratory sensitivity estimates varied in direction across estimands: coagulation-enhanced matching yielded OR 1.50 (95% CI 0.53-4.21), full-cohort logistic regression yielded adjusted OR 3.19 (95% CI 1.12-9.11), approximate adjusted time-dependent Cox analysis yielded HR 1.73 (95% CI 0.85-3.51), and strict verified-date adjusted time-dependent Cox analysis yielded HR 2.67 (95% CI 1.38-5.16).

CONCLUSIONS: The attenuation of the crude mortality difference after propensity score matching is consistent with baseline severity and confounding by indication accounting for a substantial component of the unadjusted association; however, residual confounding cannot be excluded. Because residual confounding, immortal-time bias, related time-dependent exposure bias, and proportional-hazards violations remain important limitations, these data do not establish a definitive beneficial or harmful effect of plasma transfusion. The non-significant matched result reflects limited statistical power (approximately 25% for a paired OR of 2.0) and should not be interpreted as evidence of no effect.

PMID:42471643 | DOI:10.1186/s12890-026-04497-4

J Orthop Surg Res. 2026 Jul 18;21(1):433. doi: 10.1186/s13018-026-07124-4.

ABSTRACT

This correspondence comments on the multicenter cohort study by Audigé et al. (Journal of Orthopaedic Surgery and Research, 2026) reporting outcomes of arthroscopic repair for partial-thickness supraspinatus tears. We highlight four aspects of the study that, in our view, would benefit from further clarification or additional discussion: (1) limited information regarding the surgical indications for tear completion, which may complicate interpretation of treatment-selection effects; (2) the absence of a technique-specific analysis of adverse events, particularly postoperative stiffness; (3) the unreported methodology used for intraoperative Ellman grade determination, the primary classification variable and the only factor significantly associated with tear completion; and (4) the potential residual influence of differential concomitant acromioplasty across tear-location groups despite statistical adjustment. Addressing these issues would further strengthen the interpretation and clinical applicability of this otherwise valuable prospective multicenter dataset.

PMID:42471638 | DOI:10.1186/s13018-026-07124-4

BMC Public Health. 2026 Jul 18. doi: 10.1186/s12889-026-28390-1. Online ahead of print.

ABSTRACT

BACKGROUND: Young men who have sex with men (MSM) and transgender women (TGW) in Southern Africa experience a disproportionate burden of HIV, compounded by stigma related to HIV status and sexual orientation or gender identity (SOGI). These stigmas undermine mental health and antiretroviral therapy (ART) adherence, yet few integrated, community-informed interventions exist in low-resource settings. This study assessed the feasibility, acceptability, and preliminary impact of a peer-led pilot intervention to reduce stigma, improve mental health, and strengthen ART adherence among young MSM and TGW living with HIV in Malawi, Zambia, and Zimbabwe.

METHODS: A mixed-methods parallel design was used, combining quantitative self-administered questionnaires with qualitative in-depth interviews. Seventy-two participants aged 18-24 years were recruited through community-based organisations, with 70 completing baseline and endline assessments. The intervention was initially designed as an 8-week programme grounded in the information-motivation-behavioural skills (IMB) model, however, it was subsequently reduced to 4 weeks due to time and logistical constraints and was delivered by trained peer facilitators under clinical supervision. Quantitative outcomes included measures of HIV-related stigma, SOGI stigma, depressive symptoms (CES-D-10), alcohol use (AUDIT), and ART adherence self-efficacy. Descriptive statistics were generated, paired sample t-tests were used for continuous variables, and McNemar’s tests assessed changes in categorical outcomes. Qualitative data were analysed thematically to explore participant experiences and perceptions of the intervention.

RESULTS: Statistically significant improvements were observed from baseline to endline, including reductions in internalised HIV stigma (p = 0.01), anticipated HIV stigma (p = 0.01), healthcare-related stigma (p < 0.001), SOGI-related harassment (p = 0.02), and depressive symptoms (p = 0.01). ART adherence self-efficacy increased significantly (p < 0.001). Although changes in sexual risk behaviours were not statistically significant, trends suggested increased condom use and reduced substance use before sex. Qualitative findings indicated enhanced self-acceptance, improved coping strategies (including reduced reliance on alcohol), increased confidence around disclosure, and strong acceptability of the peer-led, group-based intervention. However, some participants reported that sessions were delivered too quickly.

CONCLUSIONS: This pilot peer-led intervention was perceived feasible, acceptable, and showed promising improvements in stigma, mental health, and ART adherence self-efficacy among young MSM and TGW in three Southern African countries. The findings support the potential value of theory-informed, peer-delivered interventions to address HIV-related stigma and SOGI-related stigma and adherence challenges in highly marginalised populations, warranting further evaluation through larger and longer-term studies.

PMID:42471635 | DOI:10.1186/s12889-026-28390-1

Cancer Med. 2026 Jul;15(7):e72025. doi: 10.1002/cam4.72025.

ABSTRACT

BACKGROUND: Clinical and epidemiological evidence has suggested a potential association between kidney dysfunction and bladder cancer (BC). One hypothesis for this comorbidity is the presence of a common genetic etiology. However, little is known about the shared genetics and causality of this association. Thus, we aimed to investigate shared genetic architecture and the causal link between kidney dysfunction and bladder cancer.

METHODS: Leveraging summary statistics from large-scale genome-wide association studies (GWASs) conducted on European-ancestry populations on eGFRcrea (N = 1,004,040), eGFRcys (N = 460,826), BUN (N = 852,678), UACR (N = 288,649), urate (N = 547,361) and BC (N _cases = 3357, N _controls = 628,027), we conducted a large-scale genome-wide cross-trait study to determine genetic overlap, to identify shared loci, and to infer causal relationships. We evaluated genetic correlation through linkage disequilibrium score regression (LDSC). Integration of single-trait GWAS was accomplished by multi-trait analysis of GWAS (MTAG), which enabled cross-trait meta-analysis to unveil overlapping genetic loci between five kidney function traits and BC. Shared genes were further validated through colocalization analysis and transcriptome-wide association analysis. Bidirectional Mendelian Randomization (MR) was conducted to determine causal inference of kidney function traits on BC.

RESULTS: We found positive correlations between both BUN and urate and BC at the genome-wide level. A total of 157 significant overlapping genetic loci (range 7 to 72) were identified across five kidney function traits and BC. Among them, PSCA was prioritized as the strongest shared gene with support from MTAG, colocalization, and TWAS, whereas ZFHX3, TTC33, and ATP2A1 were considered candidates supported only by MTAG and TWAS. MR provided the most consistent support for a potential positive causal effect of BUN on BC, limited support for UACR, and exploratory evidence for urate.

CONCLUSIONS: Our cross-trait analysis demonstrated a shared genetic basis underlying kidney function and BC, providing novel insights into the biological functions and molecular mechanisms underlying these complex traits.

PMID:42470241 | DOI:10.1002/cam4.72025

Aust J Rural Health. 2026 Aug;34(4):e70234. doi: 10.1111/ajr.70234.

ABSTRACT

OBJECTIVE: To provide the first national analysis of the distribution and professional composition of PhD-qualified health professionals across metropolitan, regional, rural and remote Australia.

DESIGN: Cross-sectional analysis of 2021 Australian Census data.

SETTING: Australia, classified by Modified Monash Model (MM) remoteness categories.

PARTICIPANTS: Health professionals reporting a doctoral-level qualification (PhD).

MAIN OUTCOME MEASURES: Counts and proportions of PhD-qualified health professionals by MM category and occupation group; association between remoteness and PhD representation.

METHODS: Occupations were grouped into medicine, allied health, and nursing. Residential locations were mapped to MM categories. Descriptive statistics and linear regression examined distribution patterns.

RESULTS: Of 21 510 PhD-qualified health professionals, 85% lived in metropolitan areas (MM1), with 3027 (14%) in regional, rural and remote communities and only 12 (< 1%) in very remote areas (MM7). Medical practitioners comprised 61%, allied health 36% (predominantly psychology) and nursing 3%. Increasing remoteness was associated with reduced PhD representation (β = -0.0041, p = 0.011).

CONCLUSIONS: Research capacity is concentrated in metropolitan areas, yet a substantial base of PhD-qualified professionals exists in RRR communities. Targeted investment in funding, infrastructure and support could leverage this workforce to build sustainable rural health research capacity.

PMID:42470212 | DOI:10.1111/ajr.70234

Cancer Med. 2026 Jul;15(7):e72075. doi: 10.1002/cam4.72075.

ABSTRACT

BACKGROUND: Clinical guidelines recommend risk-management care for women with ≥ 20% lifetime risk of breast cancer (BC). Yet, guideline-concordant care-including conversations about BC risk with specialists-is low, and driving factors are not well understood. This study examined (a) associations between the material, psychosocial, and behavioral domains of financial constraint and BC risk conversations with genetics specialists and cancer/breast care specialists, and (b) whether competing resource demands are associated with the specialist BC conversations when controlling for insurance status (to disentangle the role of behavioral factors from material and psychological ones).

METHODS: Non-Hispanic Black or white women aged 18-74 years, with no history of cancer and ≥ 20% lifetime BC risk were recruited from online research volunteer databases, social media, and clinics. Participants completed an online survey about personal/family history, decision-making factors, and risk-management behavior. We used descriptive statistics to calculate proportions and both adjusted and unadjusted logistic regression to estimate odds ratios and 95% confidence intervals.

RESULTS: A total of 662 Black (34%) and white (66%) high-risk women comprised the analytic sample for this paper. Overall rates of specialist BC risk conversations were low: 43% had discussed their BC risk with any specialist (genetics specialist, breast specialist, and/or cancer specialist). In multivariate logistic regressions adjusted for race and age, perceived financial hardship was associated with lower odds of BC risk conversations with genetics (OR = 0.42 [95% CI: 0.25, 0.70]) and breast/cancer specialists (OR = 0.39 [95% CI: 0.24, 0.65]). Similar associations were observed for the relationships between disrupted insurance or single motherhood and specialist risk conversations.

CONCLUSIONS: Limited financial and other resources are associated with lower rates of BC risk conversations with specialists, which are critical to risk management in high-risk women. Future work should further investigate the complexities of resource limitations in BC risk management and develop mitigation strategies.

PMID:42470205 | DOI:10.1002/cam4.72075

JMIR Med Inform. 2026 Jul 17;14:e82924. doi: 10.2196/82924.

ABSTRACT

BACKGROUND: Electronic health records offer vast clinical data for health care research, but interoperability challenges often hinder comprehensive analysis. The Health Level Seven Fast Healthcare Interoperability Resources (FHIR) standard addresses these challenges, although its nested and interconnected resource format can be complex for analytics. Several tools have emerged to facilitate analytical access, either by querying FHIR servers via representational state transfer (REST) APIs or encoding resources in relational formats. However, the performance implications of these methods remain largely unexplored.

OBJECTIVE: This study aimed to benchmark the performance characteristics of different FHIR-based analytical approaches comparing REST API queries against SQL- and Spark-based big data frameworks operating on FHIR-encoded data.

METHODS: We benchmarked the FHIR-PYrate library, which interfaces with a FHIR server’s REST API, against Pathling, a library built for analytics based on Apache Spark, and Trino, a general-purpose SQL query engine. We defined and implemented multiple queries in each engine using 3 common analytics scenarios-data aggregation, counting, and extraction. Execution times were measured across Synthea-generated datasets of increasing size.

RESULTS: On the largest dataset, containing 71,285,064 FHIR resources, Trino completed the aggregate query more than 12,000 times faster, and Pathling did so approximately 500 times faster than FHIR-PYrate. On average across all queries, Trino outperformed FHIR-PYrate, executing extraction queries 33 times faster and count queries 1.8 times faster. Pathling achieved a 2.6-time speedup for extraction queries, but FHIR-PYrate was approximately 13 times faster for count queries.

CONCLUSIONS: While the REST-based FHIR search API is useful for standard queries and retrieving specific patient records and can outperform alternatives for some count queries, it generally lacks the performance and expressiveness needed for complex analytics. In contrast, alternative engines such as Trino and Pathling demonstrated substantial performance advantages for these scenarios.

PMID:42470189 | DOI:10.2196/82924

JMIR Rehabil Assist Technol. 2026 Jul 17;13:e88808. doi: 10.2196/88808.

ABSTRACT

BACKGROUND: Stroke is a leading cause of disability and death, resulting in a clinical, social, and economic burden upon families and the health system worldwide.

OBJECTIVE: This study aimed to introduce nurse-led rehabilitation services for patients after stroke to improve functional independence for self-care management.

METHODS: A pilot, open-label, 2-arm (1:1), randomized controlled trial was conducted at the National Institute of Neuroscience & Hospital in Bangladesh between March and August 2025. A total of 64 patients with stroke were enrolled using simple randomization. In the intervention group (IG), patients received rehabilitation education and assistive devices to improve activities of daily living. The control group (CG) received usual hospital care for stroke. Data were collected through a structured questionnaire. After discharge from the hospital, study nurses enrolled patients after stroke after a face-to-face assessment and provided rehabilitation education at the hospital and monthly rehabilitative education at home for 3 months. The study outcomes were improvement of functional independence, self-efficacy, social participation and reduction of caregivers’ burden.

RESULTS: Among 64 participants, 47 (73%) completed the study. Results depict an increase in self-care capacity, self-efficacy, social involvement and a decrease in care burden in both groups from baseline to end line; nevertheless, there was no statistically significant difference between IG and CG (F_{2, 44}=0.113; P=.74). Functional Independence Measure mean scores (IG: mean 78.4, SD 37.6; CG: mean 74.3, SD 35.4; P=.70; 95% CI -25.16 to 18.14), self-efficacy (IG: mean 27.0, SD 6.9; CG: mean 26.6, SD 7.1; P=.83; 95% CI -3.6 to 4.6), social participation (IG: mean 30.4, SD 21.4; CG: mean 28.2, SD 17.2; P=.61; 95% CI -9.33 to 13.61), and care burden (IG: mean 24.9, SD 13.6; CG: mean 25.2, SD 12.6; P=.83; 95% CI -8 to 7). In qualitative analysis, we noticed enhanced self-care capacities of patients with poststroke disabilities after rehabilitative intervention, as indicated by participants’ perceptions.

CONCLUSIONS: In this study, patients’ self-care capability was improved after the intervention from baseline, and improvements were noted by using assistive devices based on patients’ perceptions and responses. The study findings demonstrated the importance of early self-care management for patients with poststroke disabilities. Future research should focus on long-term, community-integrated strategies that involve primary care, enhanced technological support, and caregiver-focused disability adjustment programs to achieve better outcomes.

TRIAL REGISTRATION: ClinicalTrials.gov NCT06786559; https://clinicaltrials.gov/search?id=NCT06786559.

PMID:42470187 | DOI:10.2196/88808