Factors associated with long-term benzodiazepine and Z-drug use across the lifespan and 5-year temporal trajectories among incident users: a Swedish nationwide register-based study.

PURPOSE: Despite being discouraged by guidelines, long-term use of benzodiazepines and related Z-drugs (BZDR) remains frequent in the real-world. An improved understanding of factors associated with the transition from new to long-term BZDR use and of temporal BZDR use trajectories is needed. We aimed to assess the proportion of long-term BZDR use (> 6 months) in incident BZDR-recipients across the lifespan; identify 5-year BZDR use trajectories; and explore individual characteristics (demographic, socioeconomic and clinical) and prescribing-related factors (pharmacological properties of the initial BZDR, prescriber's healthcare level, and concurrent dispensing of other medications) associated with long-term BZDR use and distinct trajectories. METHODS: Our nationwide register-based cohort included all BZDR-recipients in Sweden with first dispensation in 2007-2013. Trajectories of BZDR use days per year were built using group-based trajectory modelling. Cox regression and multinomial logistic regression were fitted to assess the predictors of long-term BZDR use and trajectories' membership. RESULTS: In 930,465 incident BZDR-recipients, long-term use increased with age (20.7%, 41.0%, and 57.4% in 0-17, 18-64, and ≥ 65-year-olds, respectively). Four BZDR use trajectories emerged, labelled 'discontinued', 'decreasing', 'slow decreasing' and 'maintained'. The proportion of the 'discontinued' trajectory members was the largest in all ages, but reduced from 75.0% in the youths to 39.3% in the elderly, whereas the 'maintained' increased with age from 4.6% to 36.7%. Prescribing-related factors, in particular multiple BZDRs at initiation and concurrent dispensing of other medications, were associated with increased risks of long-term (vs short-term) BZDR use and developing other trajectories (vs 'discontinued') in all age groups. CONCLUSIONS: The findings highlight the importance of raising awareness and providing support to prescribers to make evidence-based decisions on initiating and monitoring BZDR treatment across the lifespan.

Validation of a diagnostic prediction tool for colorectal cancer: a case-control replication study.

BACKGROUND: Early detection of colorectal cancer (CRC) is crucial for survival. Primary care, the first point of contact in most cases, needs supportive risk assessment tools. We aimed to replicate the Swedish Colorectal Cancer Risk Assessment Tool (SCCRAT) for non-metastatic CRC in primary care and examine if risk factor patterns depend on sex and age. METHODS: 2,920 adults diagnosed with non-metastatic CRC during the years 2015-2019 after having visited a general practitioner the year before the diagnosis were selected from the Swedish Cancer Register and matched with 11,628 controls, using the same inclusion criteria except for the CRC diagnosis. Diagnostic codes from primary care consultations were collected from a regional health care database. Positive predictive values (PPVs) were estimated for the same 5 symptoms and combinations thereof as in the baseline study. RESULTS: The results for patients aged ≥50 years old in the present study were consistent with the results of the SCCRAT study. All symptoms and combinations thereof with a PPV >5% in the present study had a PPV >5% in the baseline study. The combination of bleeding with abdominal pain (PPV 9.9%) and bleeding with change in bowel habit (PPV 7.8%) were the highest observed PPVs in both studies. Similar risk patterns were seen for all ages and when men and women were studied separately. CONCLUSION: This external validation of the SCCRAT for non-metastatic CRC in primary care replicated the baseline study successfully and identified patients at high risk for CRC.

Drug-disease interactions in Swedish senior primary care patients were dominated by non-steroid anti-inflammatory drugs and hypertension - a population-based registry study.

OBJECTIVE: Drug-disease interactions (DDSIs) are present when a drug prescribed for one disease worsens a concomitant disease. The prevalence of DDSIs in older patients in primary care is largely unknown, as well as to what extent physicians individualize drug prescribing in relation to concomitant diseases. We therefore analysed the prevalence of DDSIs in older patients in primary care and explored to what extent physicians take possible DDSIs into account when prescribing. Design and Setting: Cross-sectional population-based register study in primary care in Region Stockholm, Sweden. Thirty-one DDSIs derived from Irish STOPP-START-Criteria were assessed. We derived data from a regional administrative healthcare database including information on all healthcare consultations and dispensed prescription drugs in the region. Data on demography, diagnoses, drug dispensations and healthcare consumption were extracted. Drugs were assessed during 2016. SUBJECTS: A total of 336,295 patients aged ≥65 registered with one of the 206 primary care practices in Region Stockholm. MAIN OUTCOME MEASURES: Prevalence and prevalence differences for DDSIs. RESULTS: In 10.8% of older patients, at least one DDSI was observed. Non-steroidal anti-inflammatory drugs (NSAIDs) were implicated in more than 75% of cases. The most common DDSI was NSAID/hypertension (8.1%), followed by NSAID/cardiovascular disease and loop diuretics/urinary incontinence (both 0.7%). The use of NSAIDs among patients with heart failure or impaired renal function was 15% lower than among patients without these diseases. CONCLUSION: DDSIs were present in every tenth older patient in primary care. Patients with cardiovascular disease receive NSAIDs to a lower extent, possibly indicating physician awareness of DDSI. Key points Evidence on the prevalence of drug-disease interactions in older patients in primary care is sparse despite their potential to cause harm. In this study, we found that every 10th older patient attending primary care had at least one drug-disease interaction. Interactions with NSAIDs were far more common than interactions with other drugs. The use of NSAIDs among patients with heart failure or impaired renal function was 15% lower than among patients without these diseases.

Secondary prevention after stroke/transient ischemic attack: A randomized audit and feedback trial.

OBJECTIVES: The clinical benefits of use of secondary preventive pharmacotherapy in ischemic stroke/TIA have been previously demonstrated. A potential target for facilitating the use of recommended medications is primary care physicians. Therefore, we carried out an audit and feedback intervention aimed at primary care centers. The aim was to improve the use of secondary preventive stroke medications and diagnosis recording in ischemic stroke/TIA. MATERIALS AND METHODS: The intervention consisted of structured, healthcare database-derived quality reports on secondary preventive medication use and diagnosis recording, sent in 2015 to half of the primary care centers in Stockholm County, with information specific to each primary care center. Medication dispensation (primary outcome) for statins, antihypertensives, antiplatelets, and anticoagulants, as well as diagnosis recording (secondary outcome), was compared between intervention centers and control centers in the 18 months following the intervention. Outcome data were derived from the healthcare databases of Stockholm County (VAL). RESULTS: Dispensation of medications to the 12 766 patients analyzed in the study was high. Over 77% of patients used antihypertensives and antithrombotics, and 65%-68% used statins. After the intervention, no differences in medication dispensation were seen between the intervention and control centers, even after adjusting for potential confounders. CONCLUSIONS: A simple audit and feedback intervention directed toward physicians in primary care did not improve medication dispensation to ischemic stroke/TIA patients 18 months later. Any future audit and feedback intervention aimed at improving adherence to guidelines for secondary prevention in primary care should consider multiple and continuous reminders, the graphical appeal, and widening the recipients to include patients.

Medication persistence to antihypertensive drug treatment - a cross-sectional study of attitudes towards hypertension and medication in persistent and non-persistent patients.

Purpose: To study the differences in attitudes towards hypertension and drug treatment between patients persistent and non-persistent to antihypertensive drug treatment. Materials and methods: Cross-sectional study on patients with hypertension treated at 25 primary healthcare centres in Stockholm, Sweden. Questionnaires were sent to the patients 3-12 months after initiation of antihypertensive drug treatment. Persistent medication users, defined as patients with less than 30 days without tablet supply between prescription refills, were compared with non-persistent users by scores from Likert scales: Brief-Illness Perception Questionnaire (Brief IPQ, 0-10) and Beliefs about Medicines Questionnaire (BMQ General, 4-20 and BMQ Specific, 5-25). Results: A total of 711 patients were included in the final analyses (mean age: 62 years; 50% women), of whom 609 (86%) were classified as persistent and 102 (14%) as non-persistent by analyses of their filled prescriptions. Likert scales from the Brief-IPQ showed (all p < 0.02) that persistent patients believed that hypertension was a chronic condition (median 6 vs. 4), that hypertension had less consequences on their life (median 2 vs. 3) and that they can prevent cardiovascular disease by taking antihypertensive treatment (median 7 vs. 5). Likert scales from the BMQ General showed (all p < 0.02) that persistent patients believed that there are potential benefits from taking the treatment (median 16 vs. 16), and they did not believe that the doctors put too much trust in drugs (median 12 vs. 13). Further, results from the BMQ Specific showed that they believed that the antihypertensive drugs are necessary for them in order to maintain or improve their own health (median 17 vs. 16). Conclusions: Primary healthcare providers should further emphasize the chronicity of hypertension diagnosis and the benefits of treatment, to improve the patients' medication persistence to antihypertensive treatment.

Drug adherence in treatment resistant and in controlled hypertension-Results from the Swedish Primary Care Cardiovascular Database (SPCCD).

PURPOSE: To assess drug adherence in patients treated with ≥3 antihypertensive drug classes, with both controlled and uncontrolled blood pressure and describe associated factors for nonadherence. METHODS: Patients with hypertension, without cardiovascular comorbidity, aged >30 years treated with ≥3 antihypertensive drug classes were followed for 2 years. Both patients with treatment resistant hypertension (TRH) and patients with controlled hypertension were included. Clinical data were derived from a primary care database. Pharmacy refill data from the Swedish Prescribed drug registry was used to calculate proportion of days covered (PDC). Patients with a PDC level ≥ 80% were included. RESULTS: We found 5846 patients treated ≥3 antihypertensive drug classes, 3508 with TRH (blood pressure ≥ 140/90), and 2338 with controlled blood pressure (<140/90 mm Hg). TRH patients were older (69.1 vs 65.8 years, P < .0001) but had less diabetes (28.5 vs 31.7%, P < .009) compared with patients with controlled blood pressure. The proportion of patients with PDC ≥ 80% declined with 11% during the first year in both groups. Having diabetes was associated with staying adherent at 1 year (RR 0.82; 95% CI, 0.68-0.98) whilst being born outside Europe was associated with nonadherence at one and (RR 2.05; 95% CI, 1.49-2.82). CONCLUSIONS: Patients with multiple antihypertensive drug therapy had similar decline in adherence over time regardless of initial blood pressure control. Diabetes was associated with better adherence, which may imply that the structured caregiving of these patients enhances antihypertensive drug treatment.

Sociodemographic factors are associated with utilisation of statins after ischaemic stroke/TIA.

OBJECTIVES: To analyse if there are sociodemographic differences in the utilisation of statins 9-12 months after ischaemic stroke or transitory ischaemic attack. METHODS: Anonymised linkage of registry data on all patients >18 years discharged from the hospitals in Stockholm, Sweden 2006-2010 with diagnosis of ischaemic stroke (ICD-10: I63.0-9) or TIA (ICD-10: G45.9) was performed. Data on hospitalisations and diagnoses were collected from the Stockholm County Council administrative databases on healthcare consumption. Dispensed prescriptions with statins and, for comparative purposes, antihypertensive agents 9-12 months after discharge were acquired from the National Swedish Prescribed Drug Register. Data about socioeconomic factors were obtained from Statistics Sweden. The dispensing of statins and antihypertensive agents, relative to sociodemographic variables were analysed. Using logistic regression odds ratios, crude, and adjusted with education, income, origin of birth, age, and sex as predictors where calculated. RESULTS: Of 24 312 patients with ischaemic stroke/TIA, 19 335 were alive 12 months after discharge. Statins were dispensed to 44% of all patients in the cohort, more frequently to men of all age groups, to patients with higher education, and to those with higher income. Antihypertensive agents were dispensed to 68% of all patients and there were no differences related to sex and income whilst patients with lower education were dispensed with antihypertensives more frequently. CONCLUSION: We find a low utilisation of statins one year after ischaemic stroke/TIA. Patients with low education, low income, and female sex were dispensed fewer prescriptions of statins indicating a need for improvement.

Non-steroidal anti-inflammatory drugs and blood pressure control in patients treated for hypertension: results from the Swedish primary care cardiovascular database.

PURPOSE: The aim of this observational cohort study was to investigate blood pressure level and the possibility to reach target blood pressure during concomitant use of NSAID in hypertensive patients. MATERIALS AND METHODS: From the Swedish primary care cardiovascular database (SPCCD) a cohort of 5463 patients (2007 to 2008) with at least one prescription of NSAID dispensed 6 months prior to the last blood pressure measurement were included. Clinical data were extracted from computerized medical records and linked to the Prescribed Drug Register. Multivariable logistic regression models were used for analysis. RESULTS: Patients with NSAID usage were younger, more often female, with lower creatinine concentrations, more musculoskeletal diagnosis and less cardiovascular comorbidity compared to patients without dispensed NSAID (p < .0001 for all). Regular dose of NSAID was not associated with a decreased possibility to reach target blood pressure. A correlation between the dose of naproxen and an increase in SBP of 7 mm Hg was found. Impairment in renal function did not influence the association between blood pressure control and NSAID (p = .27). CONCLUSION: In hypertensive patients with concomitant use of NSAID the chance to reach target blood pressure was not impaired. In intermediate and frequent users of NSAID there was a dose response relation with naproxen and SBP which was not found in diclofenac and ibuprofen.

Risks for stroke and bleeding with warfarin or aspirin treatment in patients with atrial fibrillation at different CHA(2)DS(2)VASc scores: experience from the Stockholm region.

PURPOSE: This study evaluated the benefits of and possible contraindications to warfarin treatment in patients with atrial fibrillation (AF) prior to the introduction of new oral anticoagulants using health registry data from inpatient care, specialist ambulatory care, and primary care. METHODS: This is a cohort study including all patients in the region of Stockholm, Sweden (2.1 million inhabitants) with a diagnosis of non-valvular AF (n = 41 810) recorded during 2005-2009. The risks of suffering ischemic stroke, bleeding, or death with warfarin, aspirin, or no antithrombotic treatment during 2010 were related to CHA2DS2VASc scores, age, and complicating co-morbidities. RESULTS: One-year risks for ischemic stroke were 1.0-1.2 % with aspirin, 0-0.3 % with warfarin, and 0.1-0.2 % without treatment at CHA2DS2VASc scores 0-1. Among the aspirin-treated patients with CHA2DS2VASc scores ≥2, half had possible contraindications and high risks for ischemic stroke (5.2 %), bleeding (5.0 %), and death (19.3 %). The other half of the patients with no identified contraindications had a high risk for ischemic stroke (4.0 %) but a low bleeding risk (1.8 %) and a moderate mortality rate (8.4 %). CONCLUSIONS: The present observations confirm earlier findings of undertreatment with warfarin and half of the high-risk patients treated with aspirin were obvious candidates for anticoagulant treatment. However, the other half of the patients had complicating co-morbidities, high bleeding risk, and poor prognosis. This and possible overtreatment of low-risk patients should be taken into account when considering more aggressive use of anticoagulant treatment.

The Swedish Primary Care Cardiovascular Database (SPCCD): 74 751 hypertensive primary care patients.

OBJECTIVE: To describe the Swedish Primary Care Cardiovascular Database, SPCCD. Design. Longitudinal data from electronic medical records, linked to national registers. Setting. 48 primary healthcare centres in urban (south-western Stockholm) and rural (Skaraborg) regions in Sweden. Subjects. Patients diagnosed with hypertension 2001-2008. MAIN OUTCOME MEASURES: Blood pressure (BP) and impact of retrieval of data on BP levels, clinical characteristics, co-morbidity and pharmacological treatment. RESULTS: The SPCCD contains 74 751 individuals, 56% women. Completeness of data ranged from > 99% for drug prescriptions to 34% for smoking habits. BP was recorded in 98% of patients during 2001-2008 and in 63% in 2008. Mean BP based on the last recorded value in 2008 was 142 ± 17/80 ± 13 mmHg. Digit preference in BP measurements differed between the two regions, p < 0.001. Antihypertensive drugs were prescribed in primary healthcare to 88% of the patients in 2008; however, when all prescribers were included 96% purchased their drugs. Cardiovascular co-morbidity and diabetes mellitus were present in 28% and 22%, respectively. CONCLUSION: This large and representative database shows that there is room for improvement of BP control in Sweden. The SPCCD will provide a rich source for further research of hypertension and its complications.

Effects of a pharmacist-led structured medication review in primary care on drug-related problems and hospital admission rates: a randomized controlled trial.

OBJECTIVE: To determine whether a pharmacist-led medications review in primary care reduces the number of drugs and the number of drug-related problems. DESIGN: Prospective randomized controlled trial. SETTING: Liljeholmen Primary Care Centre, Stockholm, Sweden. SUBJECTS: 209 patients aged ≥ 65 years with five or more different medications. Intervention. Patients answered a questionnaire regarding medications. The pharmacist reviewed all medications (prescription, non-prescription, and herbal) regarding recommendations and renal impairment, giving advice to patients and GPs. Each patient met the pharmacist before seeing their GP. Control patients received their usual care. MAIN OUTCOME MEASURES: Drug-related problems and number of drugs. Secondary outcomes included health care utilization and self-rated health during 12 months of follow-up. RESULTS: No significant difference was seen when comparing change in drug-related problems between the groups. However, a significant decrease in drug-related problems was observed in the intervention group (from 1.73 per patient at baseline to 1.31 at follow-up, p < 0.05). The change in number of drugs was more pronounced in the intervention group (p < 0.046). Intervention group patients were not admitted to hospital on fewer occasions or for fewer days, and there was no significant difference between the two groups regarding utilization of primary care during follow-up. Self-rated health remained unchanged in the intervention group, whereas a drop (p < 0.02) was reported in the control group. This resulted in a significant difference in change in self-rated health between the groups (p < 0.047). CONCLUSIONS: The addition of a skilled pharmacist to the primary care team may contribute to reductions in numbers of drugs and maintenance of self-rated health in elderly patients with polypharmacy.

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.

Persistence to antihypertensive drug treatment in Swedish primary healthcare.

PURPOSE: To determine factors associated with low persistence in patients initiated on drug treatment for hypertension. METHODS: Cohort study using medical records for patients with hypertension in 48 Swedish primary healthcare centres. Data were linked to national registers on dispensed drugs, hospitalizations, outpatient hospital consultations, deaths, migration, and socioeconomy. We identified 5225 patients (55 % women, mean age 61 years) initiated on antihypertensive drug treatment during 2006-2007. Persistence was measured for two years by the dispensed drugs. Patients with a gap of >30 days between end of dispensed supply and the next dispensed prescription were classified as non-persistent. This was calculated by Kaplan-Meier analysis. Cox proportional hazard regression was used to estimate hazard ratios for discontinuation. Potential predictors included age, gender, blood pressure before initiation of therapy, cardiovascular comorbidity, educational level, country of birth, and income. RESULTS: Among patients with a dispensed first prescription, 26 % discontinued treatment during the first year, and a further 9 % discontinued during the second year. Discontinuation (all adjusted) was more common in men (P = 0.002) and in younger patients (30-49 years, P < 0.001). Systolic (P < 0.001) but not diastolic blood pressure was positively associated with persistence. Native-born Swedish citizens and patients born in the other Nordic countries had lower discontinuation rates than those born outside the Nordic countries (P < 0.001). CONCLUSION: Major determinants of discontinuation of antihypertensive drug treatment are male sex, young age, mild-to-moderate systolic blood pressure elevation, and birth outside of Sweden.

Lack of long-term effects from a pedagogical intervention upon blood pressure control--a randomized primary care study.

OBJECTIVE: The primary aim was to evaluate the effect on blood pressure (BP) levels of a previously developed pedagogically structured BP card introduced to patients with inadequately controlled hypertension in primary care. The evaluation was based on the results of a pilot study which is briefly presented. The aim of the study was to validate the positive results from a pilot study in a different, larger setting, for a longer time, and to study the effects of a nurse-led individual health counseling strategy. DESIGN: A "BP card" that summarized the essential targets of hypertension treatment was presented to patients with a small set of questions. An open, randomized, controlled study was performed testing the effect of the BP card: BP card with an added semi-structured nurse counseling versus usual care (3 groups) during 12 months. RESULTS: The effects on BP levels differed greatly from results seen in the pilot study where BP fell significantly in the intervention group as compared with that in the control group. In the main study, however, BP levels declined more than 25/8.5 mmHg in all three groups, with no significant differences between the groups. CONCLUSION: The positive results in the pilot study could not be confirmed in the main study. Furthermore, the nurse-led individual health counseling strategy did not show any additive effects. The reasons for these discrepant findings may be external such as increased awareness of hypertension, and internal factors such as contamination and non-biased recruitment.

GPs' experiences of phosphatidylethanol in treatment of hypertension: a qualitative study.

BACKGROUND: Hazardous alcohol use increases the risk of hypertension but is underdetected in primary healthcare (PHC) patients. Use of the biomarker phosphatidylethanol (PEth), which reflects the last 2-3 weeks of alcohol consumption, is increasing in Swedish PHC, but studies exploring its use for hypertension are scarce or missing. AIM: To explore GPs' experiences of using PEth to identify hazardous alcohol use in the context of managing hypertension. DESIGN & SETTING: A qualitative study of GPs (n = 12) experienced in using PEth in hypertension management who were recruited at Swedish primary healthcare centres (PHCC) in 2021. METHOD: The GPs participated in five focus group interviews. A questioning route was used. The interviews were audio-recorded, transcribed verbatim, and analysed with inductive qualitative content analysis. RESULTS: 'I don't hesitate anymore' was the overall theme, which reflected both the disappearance of GPs' fear that the PEth result might upset the patient, as this rarely occurred, and that the positive effects of PEth predominated in the findings. The theme is underpinned by the following four sub-themes: serving as an eye-opener; improving the dialogue; using with care; and learning by doing. CONCLUSION: PEth is a useful tool that changed GPs' routines for addressing alcohol and identifying hazardous alcohol use in patients with hypertension managed in PHC. The GPs advocated adopting PEth as a routine test in the treatment of hypertension. However, PEth needs to be used with care to maximise benefit and minimise harm.

Physical-mental multimorbidity in a large primary health care population in Stockholm County, Sweden.

BACKGROUND: Multimorbidity of mental and physical disorders may be common. Our objectives were to estimate the prevalence of mental disorders and determine physical-mental multimorbidity relationships adjusted for sex and age within a primary care population in Region Stockholm, Sweden. METHODS: From 2.4 million Region Stockholm inhabitants, we included adult patients with ≥ 1 primary care consultation from 2013 through 2017. We clustered 40 physical diagnoses into 9 categories and grouped mental disorders into mild-moderate (i.e., depression, anxiety, stress disorder, sleep disturbance) or severe (i.e., bipolar disorder, schizophrenia). RESULTS: Of 1 105 065 patients, mean age was 49 years, 56% were females, and nearly one-third had a mental disorder (97% mild-moderate). Adjusted odds ratios (AOR) for mild-moderate and severe mental disorders were highest in patients with alcohol abuse (AOR=3.7, 95% CI 3.6-3.8; AOR=7.2, 95% CI 6.7-7.6, respectively) vs. those with no abuse. Higher odds for either level of mental comorbidity occurred in patients with chronic heart failure (CHF), cerebrovascular disease, Transient ischemic attack (TIA), Chronic obstructive pulmonary disease (COPD), Gastroesophageal reflux disease- irritable bowel syndrome (GERD-IBS), chronic pain, dementia, nicotine dependence, and Parkinson's disease. For mild-moderate mental disorders, AOR in males was highest (1.45) at age 28 and was below 1.0 after age 46; AOR in females was highest (1.30) at age 38 and was below 1.0 after age 38. For severe mental disorders, AOR was below 1.0 after age 58 in males and after age 62 in females. CONCLUSION: Physical-mental multimorbidity was common in primary care patients in Sweden, with the highest odds occurring in those who were female, younger, and/or had an alcohol abuse disorder.

Correction: Lung cancer prediction using machine learning on data from a symptom e-questionnaire for never smokers, formers smokers and current smokers.

[This corrects the article DOI: 10.1371/journal.pone.0276703.].

A machine learning tool for identifying non-metastatic colorectal cancer in primary care.

BACKGROUND: Primary health care (PHC) is often the first point of contact when diagnosing colorectal cancer (CRC). Human limitations in processing large amounts of information warrant the use of machine learning as a diagnostic prediction tool for CRC. AIM: To develop a predictive model for identifying non-metastatic CRC (NMCRC) among PHC patients using diagnostic data analysed with machine learning. DESIGN AND SETTING: A case-control study containing data on PHC visits for 542 patients >18 years old diagnosed with NMCRC in the Västra Götaland Region, Sweden, during 2011, and 2,139 matched controls. METHOD: Stochastic gradient boosting (SGB) was used to construct a model for predicting the presence of NMCRC based on diagnostic codes from PHC consultations during the year before the date of cancer diagnosis and the total number of consultations. Variables with a normalised relative influence (NRI) >1% were considered having an important contribution to the model. Risks of having NMCRC were calculated using odds ratios of marginal effects. RESULTS: Of the 361 variables used as predictors in the stochastic gradient boosting model, 184 had non-zero influence, with 16 variables having NRI >1% and a combined NRI of 63.3%. Variables representing anaemia and bleeding had a combined NRI of 27.6%. The model had a sensitivity of 73.3% and a specificity of 83.5%. Change in bowel habit had the highest odds ratios of marginal effects at 28.8. CONCLUSION: Machine learning is useful for identifying variables of importance for predicting NMCRC in PHC. Malignant diagnoses may be hidden behind benign symptoms such as haemorrhoids.

A Qualitative Study on Patients' Views on Hypertension and Antihypertensive Medications.

Purpose: Few studies explored what patients initiated on blood pressure medication in primary care think about their disease and their medications. The aim of this study was to gain an understanding of hypertensive patients' views on and experiences of hypertension and the use of antihypertensive medications. Methods: A qualitative study based on open-ended questions from a survey on medication adherence, which captured treated hypertensive patients' perspective on their condition and treatment. Data were collected for 219 patients on antihypertensive medication, ≥ 30 years old, who consulted 25 primary health care centers in Stockholm, Sweden, during 2016. Thematic analysis with both inductive and deductive approach was applied. Results: We identified 21 codes from the data and grouped them under the World Health Organization's five dimensions of adherence: condition-, therapy-, health care team and system-, patient-, and socioeconomic-related factors. The analyses revealed that many patients with hypertension have limited knowledge of their disease, are afraid of drug side effects and experience various issues in primary health care that may negatively impact adherence, including short doctor appointments, prescribing without communication and room for improvement in individualization of therapy and a person-centered approach. Conclusion: Many patients with hypertension have limited understanding of their hypertension and fear of adverse events from their antihypertensive medications. There is also room for improvement in how the patients are managed in primary health care. Interventions should focus on these issues to promote a better blood pressure target achievement.

Lung cancer prediction using machine learning on data from a symptom e-questionnaire for never smokers, formers smokers and current smokers.

PURPOSE: The aim of the present study was to investigate the predictive ability for lung cancer of symptoms reported in an adaptive e-questionnaire, separately for never smokers, former smokers, and current smokers. PATIENTS AND METHODS: Consecutive patients referred for suspected lung cancer were recruited between September 2014 and November 2015 from the lung clinic at the Karolinska University Hospital, Stockholm, Sweden. A total of 504 patients were later diagnosed with lung cancer (n = 310) or no cancer (n = 194). All participants answered an adaptive e-questionnaire with a maximum of 342 items, covering background variables and symptoms/sensations suspected to be associated with lung cancer. Stochastic gradient boosting, stratified on smoking status, was used to train and test a model for predicting the presence of lung cancer. RESULTS: Among never smokers, 17 predictors contributed to predicting lung cancer with 82% of the patients being correctly classified, compared with 26 predictors with an accuracy of 77% among current smokers and 36 predictors with an accuracy of 63% among former smokers. Age, sex, and education level were the most important predictors in all models. CONCLUSION: Methods or tools to assess the likelihood of lung cancer based on smoking status and to prioritize investigative and treatment measures among all patients seeking care with diffuse symptoms are much needed. Our study presents risk assessment models for patients with different smoking status that may be developed into clinical risk assessment tools that can help clinicians in assessing a patient's risk of having lung cancer.