Predicting new cases of hypertension in Swedish primary care with a machine learning tool.

Background: Many individuals with hypertension remain undiagnosed. We aimed to develop a predictive model for hypertension using diagnostic codes from prevailing electronic medical records in Swedish primary care. Methods: This sex- and age-matched case-control (1:5) study included patients aged 30-65 years living in the Stockholm Region, Sweden, with a newly recorded diagnosis of hypertension during 2010-19 (cases) and individuals without a recorded hypertension diagnosis during 2010-19 (controls), in total 507,618 individuals. Patients with diagnoses of cardiovascular diseases or diabetes were excluded. A stochastic gradient boosting machine learning model was constructed using the 1,309 most registered ICD-10 codes from primary care for three years prior the hypertension diagnosis. Results: The model showed an area under the curve (95 % confidence interval) of 0.748 (0.742-0.753) for females and 0.745 (0.740-0.751) for males for predicting diagnosis of hypertension within three years. The sensitivity was 63 % and 68 %, and the specificity 76 % and 73 %, for females and males, respectively. The 25 diagnoses that contributed the most to the model for females and males all exhibited a normalized relative influence >1 %. The codes contributing most to the model, all with an odds ratio of marginal effects >1 for both sexes, were dyslipidaemia, obesity, and encountering health services in other circumstances. Conclusions: This machine learning model, using prevailing recorded diagnoses within primary health care, may contribute to the identification of patients at risk of unrecognized hypertension. The added value of this predictive model beyond information of blood pressure warrants further study.

Diagnostic Guidelines for Familial Hemophagocytic Lymphohistiocytosis Revisited.

Current HLH-2004-based diagnostic criteria for familial hemophagocytic lymphohistiocytosis (FHL) are based on expert opinion. Here we performed a case-control study to test and possibly improve these clinical criteria. We also developed two complementary expert opinion-based diagnostic strategies for FHL in patients with signs/symptoms suggestive of HLH, based on genetic and cellular cytotoxicity assays. The cases (n=366) were children <16 years with verified familial and/or genetic FHL (n=341) or Griscelli syndrome type 2 (GS2) (n=25); 276 from the HLH-94/HLH-2004 databases and 90 from the Italian HLH Registry. All fulfilled the HLH-94/HLH-2004 patient inclusion criteria. Controls were 374 children with systemic-onset juvenile idiopathic arthritis (sJIA) and 329+361 children in two cohorts with febrile infections that could be confused with HLH and sepsis, respectively. To provide complete data sets, multiple imputations were performed. The optimal model, based on the number of diagnostic criteria fulfilled from 17 variables studied, reveled almost similar diagnostic thresholds as the existing criteria, with accuracy 99.1% (sensitivity 97.1%; specificity 99.5%). Notably, assessment of the original HLH-2004 criteria revealed accuracy 97.4% (sensitivity 99.0%; specificity 97.1%). Since cellular cytotoxicity assays here constitute a separate diagnostic strategy, HLH-2004 criteria without NK-cell function was also studied which showed accuracy 99.0% (sensitivity 96.2%; specificity 99.5%). Thus, we conclude that the HLH-2004 criteria (without NK-cell function) have significant validity in their current form when tested against severe infections or sJIA. It is important to exclude underlying malignancies and atypical infections. In addition, complementary cellular and genetic diagnostic guidelines can facilitate necessary confirmation of clinical diagnosis.

Changes in Sleep Quality, Sleep Duration, and Sickness Absence: A Longitudinal Study with Repeated Measures.

BACKGROUND: Sickness absence has been linked to short and long, as well as poor, sleep in a few studies. Such studies have started from a baseline measurement and followed up on subsequent sickness absence. In the present study, however, we focused on the change in biennial reports of sickness absence and sleep measures (using work-related variables as possible modifiers). We also searched for an interaction between predictors and gender since women report more sleep problems. METHODS: A total of 5377 individuals (random sample from the Swedish working population) participated across five biennial points of measurement. Data were analyzed using mixed-model logistic regression. RESULTS: The multivariable analysis of variation across the five time points showed that the significant sleep-related predictors of sickness absence (at least one occurrence during the preceding year) were sleep duration during days off (OR = 1.16, 95% Cl = 1.08;1.24) and sleep problems (OR = 1.42, 95% CI = 1.33;1.51). These also remained significant after the addition of psychosocial work factors. Sensitivity analyses indicated that a 9 h sleep duration during days off may represent a critical level in terms of increased sickness absence and that late rising contributed to the association between sickness absence and long sleep duration during days off. Women reported a higher sickness absence than men (OR = 2.16, 95% CI = 1.74;2.68) and had a higher probability of sickness absence for long sleep during days off and during the workweek than men. CONCLUSIONS: It was concluded that increases in sleep problems and sleep duration during days off are longitudinally associated with changes in sickness absence and that women have a closer link between the two. This suggests that treatment for sleep problems may reduce the risk of sickness absence.

A machine learning tool for identifying patients with newly diagnosed diabetes in primary care.

BACKGROUND AND AIM: It is crucial to identify a diabetes diagnosis early. Create a predictive model utilizing machine learning (ML) to identify new cases of diabetes in primary health care (PHC). METHODS: A case-control study utilizing data on PHC visits for sex-, age, and PHC-matched controls. Stochastic gradient boosting was used to construct a model for predicting cases of diabetes based on diagnostic codes from PHC consultations during the year before index (diagnosis) date and number of consultations. Variable importance was estimated using the normalized relative influence (NRI) score. Risks of having diabetes were calculated using odds ratios of marginal effects (ORME). Four groups by age and sex were studied, age-groups 35-64 years and ≥ 65 years in men and women, respectively. RESULTS: The most important predictive factors were hypertension with NRI 21.4-29.7 %, and obesity 4.8-15.2 %. The NRI for other top ten diagnoses and administrative codes generally ranged 1.0-4.2 %. CONCLUSIONS: Our data confirm the known risk patterns for predicting a new diagnosis of diabetes, and the need to test blood glucose frequently. To assess the full potential of ML for risk prediction purposes in clinical practice, future studies could include clinical data on life-style patterns, laboratory tests and prescribed medication.

Colonoscopy findings after increasing two-stool faecal immunochemical test (FIT) cut-off: Cross-sectional analysis of the SCREESCO randomized trial.

BACKGROUND: We determined the impact of an increased two-stool faecal immunochemical test (FIT) cut-off on colonoscopy positivity and relative sensitivity and specificity in the randomized controlled screening trial screening of Swedish colons conducted in Sweden. METHODS: We performed a cross-sectional analysis of participants in the FIT arm that performed FIT between March 2014 and 2020 within the study registered with ClinicalTrials.gov, NCT02078804, who had a faecal haemoglobin concentration of at least 10 µg/g in at least one of two stool samples and who underwent a colonoscopy (n = 3841). For each increase in cut-off, we computed the positive predictive value (PPV), numbers needed to scope (NNS), sensitivity and specificity for finding colorectal cancer (CRC) and advanced neoplasia (AN; advanced adenoma or CRC) relative to cut-off 10 µg/g. RESULTS: The PPV for AN increased from 23.0% (95% confidence intervals [CI]: 22.3%-23.6%) at cut-off 10 µg/g to 28.8% (95% CI: 27.8%-29.7%) and 33.1% (95% CI: 31.9%-34.4%) at cut-offs 20 and 40 µg/g, respectively, whereas the NNS to find a CRC correspondingly decreased from 41 to 27 and 19. The PPV for AN was higher in men than women at each cut-off, for example 31.5% (95% CI: 30.1%-32.8%) in men and 25.6% (95% CI: 24.3%-27.0%) in women at 20 µg/g. The relative sensitivity and relative specificity were similar in men and women at each cut-off. CONCLUSION: A low cut-off of around 20-40 µg/g allows detection and removal of many AN compared to 10 µg/g while reducing the number of colonoscopies in both men and women.

Lifestyle counseling in patients with hypertension in primary health care and its association with antihypertensive pharmacotherapy.

The study aimed to investigate differences in hypertensive- and cardio-preventive pharmacotherapy depending on if patients with hypertension received lifestyle counseling or not, including the difference between men and women. Data from the Region Stockholm VAL database was used to identify all patients with a hypertension diagnosis and had visited a primary health care center within the past five years. Data included registered diagnoses, pharmacotherapy, and codes for lifestyle counseling. Logistic regression adjusted for age and comorbidity (diabetes, stroke, coronary heart disease, atrial fibrillation, gout, obesity, heart failure) was used, presenting results as odds ratios (OR) with 99% confidence interval (CI). The study included 130,030 patients with hypertension; 63,402 men and 66,628 women. Patients receiving recommended lifestyle counseling were more frequently treated with three or more hypertensive drugs: women OR 1.38 (1.31, 1.45) and men = 1.36 (1.30, 1.43); certain drug classes: calcium antagonists: women 1.09 (1.04, 1.14) and men 1.11 (1.06, 1.16); thiazide diuretics: women 1.26 (1.20, 1.34) and men 1.25 (1.19, 1.32); and aldosterone antagonists: women 1.25 (1.12, 1.41) and men 1.49 (1.34, 1.65). Patients receiving recommended level of lifestyle counseling with concomitant coronary heart disease, atrial fibrillation, diabetes, or stroke were more frequently treated with statins than those who did not. Further, recommended lifestyle counseling was significantly associated with anticoagulant treatment in patients with atrial fibrillation. Lifestyle counseling according to recommendations in national guidelines was significantly associated with a more thorough pharmacological treatment of hypertension, statins, and antithrombotic drugs as well as anticoagulants, in both men and women.

The association of short and long sleep with mortality in men and women.

Both short (< 6 hr) and long (> 8 hr) sleep are associated with increased mortality. We here investigated whether the association between sleep duration and all-cause, cardiovascular disease and cancer mortality differs between men and women. A cohort of 34,311 participants (mean age and standard deviation = 50.5 ± 15.5 years, 65% women), with detailed assessment of sleep at baseline and up to 20.5 years of follow-up (18 years for cause-specific mortality), was analysed using Cox proportional hazards model to estimate HRs with 95% confidence intervals. After adjustment for covariates, all-cause, cardiovascular disease and cancer mortalities were increased for both < 5 hr and ≥ 9 hr sleep durations (with 6 hr as reference). For all-cause mortality, women who slept < 5 hr had a hazard ratio = 1.54 (95% confidence interval = 1.32-1.80), while the corresponding hazard ratio was 1.05 (95% confidence interval = 0.88-1.27) for men, the interaction being significant (p < 0.05). For cardiovascular disease mortality, exclusion of the first 2 years of exposure, as well as competing risk analysis eliminated the originally significant interaction. Cancer mortality did not show any significant interaction. Survival analysis of the difference between the reference duration (6 hr) and the short duration (< 5 hr) during follow-up showed a gradually steeper reduction of survival time for women than for men for all-cause mortality. We also observed that the lowest cancer mortality appeared for the 5-hr sleep duration. In conclusion, the pattern of association between short sleep duration and all-cause mortality differed between women and men, and the difference between men and women increased with follow-up time.

Intimate partner violence and negative health consequences: A cross-sectional study among women in a regional sample in Sweden.

BACKGROUND: Intimate partner violence (IPV) is a global health problem of enormous proportions. However, little is known about the prevalence or health consequences of IPV among women in Stockholm, Sweden, a city characterised by high levels of gender equality that hosts a large population of people born outside Europe. AIMS: This study aimed to assess the prevalence of exposure to physical, psychological and sexual IPV and its associated background factors and health outcomes. METHODS: This was a cross-sectional study employing a survey containing questions about the previous year's exposure to IPV that was distributed to 35 midwifery clinics in Stockholm during the autumn of 2020. Any woman who visited any of these midwifery clinics during these two months was eligible to participate. RESULTS: A total of 2239 women answered the questionnaire, of whom 25.1% reported having been subjected to IPV at some point during their life and 8.7% during the previous year. The most common ongoing exposure was psychological violence, which was reported by 6.6% of the women. Women living with IPV reported poorer self-rated general health and more recurring health symptoms and depression than unexposed women. CONCLUSIONS: Exposure to IPV is common and is associated with depression, lower general well-being and somatic health problems.

Comparison of Methods for Monitoring Intra-operative Cerebral Perfusion in Patients Undergoing Carotid Endarterectomy with Selective Shunting: A Systematic Review and Network Meta-Analysis of Randomised Controlled Trials and Cohort Studies.

OBJECTIVE: This study aimed to analyse the influence of different methods of monitoring cerebral perfusion (MCP) on stroke, death, and use of intraluminal shunt during carotid endarterectomy (CEA). METHODS: A systematic review and network meta-analysis was conducted and registered in the PROSPERO registry (CRD42021246360). Medline, Embase, CENTRAL, and Web of Science were searched. Randomised controlled trials (RCTs) and cohort studies with > 50 participants that compared clinical outcomes for different MCP in patients undergoing CEA were included. Papers reporting one or a combination of two of the following MCPs were included in the analysis: awake testing (AT), near infrared spectroscopy (NIRS), electroencephalography (EEG), somatosensory evoked potential (SSEP), motor evoked potential (MEP), transcranial Doppler (TCD), and stump pressure (SP). A random effects network meta-analysis was performed using a binomial likelihood function with a specified logit link for peri-operative stroke or death and shunting as outcomes. Near infrared spectroscopy was excluded due to the lack of studies that could be used for statistical analysis. RESULTS: Of 1 834 publications, 17 studies (15 cohort studies and two RCTs) including 21 538 participants were incorporated in the quantitative analysis. Electroencephalography was used in the largest number of participants (7 429 participants, six studies), while AT was used in the highest number of studies (10 studies). All monitoring modalities had worse outcomes with respect to stroke or death when compared with AT, with ORs ranging between 1.3 (95% credible interval [CrI] 0.2 - 10.9) for SSEP + MEP and 3.1 (CrI 0.3 - 35.0) for patients monitored with a combination of EEG and TCD. However, the wide CrI indicated that there is no statistically significant difference between the monitoring methods. Patients monitored with a combination of EEG and TCD had the lowest odds of being shunted, while SP had the highest odds of being shunted, also with no statistically significant difference. CONCLUSION: There is a lack of high quality data on this topic in the literature. The present study showed no significant difference between monitoring methods investigated in the network meta-analysis.

Effects of internet-delivered cognitive behavioral therapy on use of child sexual abuse material: A randomized placebo-controlled trial on the Darknet.

Introduction: The use of child sexual abuse material (CSAM) is an international public health and child protection challenge. Objective: To investigate whether Prevent It, a therapist-supported, internet-delivered, eight-week, cognitive behavioral therapy, reduces CSAM viewing among users. Methods: We conducted a global online single-blind (participants), parallel-group, superiority, randomized, psychological placebo-controlled trial with a one-month follow-up, 2019-2021 (ISRCTN76841676). We recruited anonymous participants, mainly from Darknet forums. Inclusion criteria: age 18+ years, past week CSAM use, and sufficient English language skills; exclusion criteria: severe psychiatric illness or non-serious intent to participate. The main outcome was change in self-reported, weekly viewing time from pre- to post-treatment, according to the Sexual Child Molestation Risk Assessment+. Results: A total of 160 participants (157 male, 2 non-binary, and 1 not reporting gender) from all world regions (age intervals [%]: 18-29 [49]; 30-39 [30]; 40-49 [15]; 50-59 [6]) were randomized (1:1) to Prevent It (N = 80) or Placebo (N = 80). Between-group, intention-to-treat analyses suggested a significantly larger decrease in viewing time in Prevent It participants vs. controls pre- to post-treatment (Prevent It: N = 76, Placebo: N = 78, estimate -0.25, 95 % CI, -0.46 to -0.04, p = .017, Cohen's d 0.18). Negative side effects from treatment were fewer in Prevent It compared to control participants and neither group reported severe adverse events. Conclusion: We provide initial support for the feasibility, efficacy, and safety of Prevent It to reduce CSAM viewing among motivated users. Further research is needed to validate these findings.

Postoperative complications and myocardial injury in patients receiving air or oxygen. Prospective, randomised and pilot study.

BACKGROUND: Supplementary oxygen is administered during anaesthesia to increase oxygen delivery and prevent hypoxia. Recent studies have questioned this routine. In this pilot study, our main aim was to investigate if 21% oxygen compared to ≥50% reduces the risk of postoperative complications and myocardial injury. METHODS: In this pragmatic, multicentre, single-blind study, patients undergoing vascular surgery were randomised to receive a fraction of inspired oxygen (Fi O2 ) ≥ 0.50 and oxygen saturation determined by pulse oximetry (SpO2 ) ≥ 98% (group H) or Fi O2 of 0.21 and SpO2 > 90% (group N) oxygen perioperatively. The primary outcome was a composite outcome of major pre-defined postoperative complications assessed at 30 days. Myocardial injury was determined by serial troponin measurements. Data were analysed using generalised estimating equation, Mann-Whitney U test or chi-squared test, as appropriate. RESULTS: The 191 patients were randomised, and per-protocol principle was used for analyses. At 30-day follow-up, 43 out of 94 patients (46%) had a postoperative complication in group H and 36 out of 90 patients (40%) in group N, p = .46. New myocardial injury was seen in 27% versus 22% in Groups H and N respectively (p = .41). No differences in other outcomes were observed between the groups. Twelve patients (13%) in Group N had SpO2 < 90%, six recovered spontaneously and six required supplemental oxygen. At 1-year follow-up, one patient in group H had died. CONCLUSION: In this pilot study, postoperative complications were similar between the groups in patients randomised to Fi O2 of 0.21 or ≥0.50 and no difference was found in the incidence of new myocardial injury. Larger, prospective adequately powered studies are needed.

Bone Loss Before and After Peri-implantitis Surgery: A 7-Year Retrospective Observational Study.

PURPOSE: This study aimed to report changes in implant bone levels before and after peri-implantitis surgery (P-IS). MATERIALS AND METHODS: Patients were treated with P-IS from 2003 to 2010 and thereafter followed up until the last available radiographic examination (study termination: October 2018). Bone loss was measured before and after P-IS, and the numbers of clinical examinations post-P-IS were determined. The Student t test was used to compare the bone loss between examinations and reference populations, and correlations between different variables were calculated using the Pearson correlation coefficient (r). RESULTS: One hundred thirty-four consecutive patients (137 arches) underwent P-IS during the inclusion period (2.3%). A total of 122 patients (125 arches) were followed up for a mean period of 6.9 (SD: 5.54) and 7.3 (SD: 3.39) years from the first annual examination to before P-IS, and from the time of P-IS to after P-IS, respectively. The mean bone loss during the aforementioned durations was 0.18 (SD: 0.23) and 0.26 (SD: 0.28) mm/year, respectively (P < .05). Bone loss per year was also greater for unaffected implants after (0.29 [0.34] mm/year) compared with before P-IS (0.06 [0.10] mm/year; P < .05). No difference in bone loss was observed between peri-implantitis affected and unaffected implants or between implants with a turned or moderately rough surface after P-IS. A greater amount of bone loss was associated with increasing numbers of clinical examinations and shorter follow-up times after P-IS (P < .05). CONCLUSION: Average bone levels decreased from the time of prosthesis placement to termination of the study. Bone loss increased after P-IS in the total and unaffected group, and edentulous patients compared unfavorably with partially and single implant conditions. Increased bone loss was associated with higher numbers of clinical examinations per year of followup after P-IS (P < .05).

Implant failures before and after peri-implantitis surgery: A retrospective study on 207 consecutively treated patients.

PURPOSE: This study aimed to report implant failures before and after peri-implantitis surgery (P-IS) and to compare the pattern of implant failures with in untreated at-risk patients during the same period. MATERIALS AND METHODS: A total of 5628 untreated at-risk patients (7240 jaws) and 207 consecutively patients treated with P-IS (212 jaws) were included in two groups. Implants were placed and followed-up between 1986 and 2018. Cumulative survival rates (CSR) for patients treated with P-IS after 2003 were calculated before and after P-IS and compared with those for nontreated patients. RESULTS: The overall 15-year CSR was 91.2% (CI 95%; 90.5%-91.9%) and 68.5%, (CI 95%, 62.1%-75.5%) for untreated patients at risk and P-IS patients, respectively (P < .05). The 10-year CSR (baseline 1 year after implant surgery) was 97.2% (CI 95%, 95.2%-100%) for treated patients before P-IS which was comparable with that for untreated patients: 95.4% (CI 95%, 94.8%-97.7%). The corresponding 10-year CSR for P-IS patients after surgery was significantly lower (71.6%: CI 95%, 63.1%-81.3% (P < .05)). CONCLUSION: CSR for patients/jaws without implant failures was comparable between untreated and treated P-IS patients before, but lower for P-IS patients after P-IS (P < .05). A negative effect of P-IS on implant survival after treatment cannot be disregarded.

Registered nurses' perceptions of safe care in overcrowded emergency departments.

AIMS AND OBJECTIVES: To explore registered nurses' perceptions of safe practice in care for patients with an extended length of stay in the emergency department. BACKGROUND: Extended length of stay and overcrowding in emergency departments are described internationally as one of the most comprehensive challenges of modern emergency care. An emergency department is not designed, equipped or staffed to provide care for prolonged periods of time. This context, combined with a high workload, poses a risk to patient safety, with additional medical errors and an increased number of adverse events. From this perspective, it is important to extend our knowledge and to describe registered nurses' experiences of safe practice. DESIGN: A qualitative, inductive and descriptive study. METHODS: Qualitative interview study carried out in five emergency departments. Data were analysed using a qualitative content analysis with a latent approach. RESULTS: Patient safety meets obstacles in the clinical environment involving experiencing deficiencies regarding patient safety in the clinical setting and the impact of working procedures and routines. Moreover, nurses are challenged in their professional responsibilities involving balancing essential nursing care and actual workload; it is common to experience emotional reactions based on feelings of loss of control. CONCLUSIONS: From the nurses' perspective, a prolonged stay in the emergency department may lead to negative consequences for both patient safety and care as well as registered nurses' psychosocial experiences. An extended length of stay significantly reduces the level of nursing and caring that registered nurses can perform in the emergency department. RELEVANCE TO CLINICAL PRACTICE: This study indicates that emergency departments should review their procedures to avoid both deviations from normal practice and moral stress among registered nurses. This can contribute to an increased understanding and insight about the challenge of patient safety in an emergency department setting.