Landelijk Trial Register (NL8041; NL72069.029.19); registered 2019-09-23; covering all WHO Trial Registration Data Set items.

The ProIntens study is a two-armed multicentre parallel individually randomised trial. After determining eligibility and obtaining informed consent, participants are randomised 1:1 to either the control group or intervention group. Participants in the control group will receive regular care (RC), while participants in the intervention group will receive intensive dietetic care (IDC). Measurements will be performed at hospital admission (t ₀), discharge (t ₂), and at one (t ₃)- and three-months after discharge (t ₄).

The study population will consist of hospitalised older adults (≥55y), who are at risk of malnutrition. Five hospitals in the region of Amsterdam will participate: Amsterdam University Medical Centers, location AMC (1000 beds) and VUmc (733 beds), teaching hospital OLVG, location Oost (555 beds) and West (366 beds) and regional hospital BovenIJ (313 beds). Participants will be recruited from the following medical wards: geriatrics, internal medicine, cardiology, gastroenterology, pulmonology and the acute admission unit (a non-emergency triage ward where patients undergo further tests and stabilisation before being transferred to a suitable ward). To identify participants at risk of malnutrition, four validated screening tools specifically designed for older adults and the hospital setting will be performed within 48 hours after hospital admission, namely: Short Nutritional Assessment Questionnaire (SNAQ), Malnutrition Universal Screening Tool (MUST), Mini Nutritional Assessment Short Form (MNA-SF), and the Malnutrition Screening Tool (MST). ¹² The cut-off value for the risk of malnutrition for SNAQ is ≥3 points, for MUST is ≥2 points, for MST is ≥2 points and for MNA-SF is ≤11 points. Participants at risk of malnutrition are eligible for inclusion if one or more of the four screening tools indicates such risk as by their respective cut-offs. The patient will be approached for participation after consultation with the clinical team. After obtaining written informed consent, the participant is further screened for eligibility. Other in- and exclusion criteria are listed in Table 1. Recruitment for the study will start 21 ^st of January 2021 and will continue until 31 ^st of December 2022.

The sample size is based on the difference in changes of the Short Physical Performance Battery (SPPB) in the study of Tieland et al ¹³ and observational data from older patients, aged 70 and older, who were at risk of malnutrition. ¹⁴ The following estimates were used: effect size of 1-point difference in mean change from admission to 3 months post-discharge between the intervention and control group and a standard deviation (SD) of 2.5. With a significance level (α) of 0.05 and a power of 80%, a sample size of 99 participants per group was estimated. Assuming a drop-out of approximately 25%, a total number of 125 participants per group was needed, leading to a total of 250 participants to be included in the study.

Eligible participants will be contacted by a member of the research team and after receiving written and oral explanation of the study procedures, the participants will have at least 3 hours to consider participation. When all questions are answered and the participant expresses his/her willingness to participate, the informed consent form will be signed, after which the participants will be screened on eligibility (see Table 1). Eligible patients will be randomized using data management program Castor EDC block randomisation algorithm (Castor EDC©, Amsterdam, The Netherlands). Variable block sizes ^{2, 4, 6} are used in random order, stratified by hospital. When a participant’s case report form is entered in Castor EDC and fulfils all inclusion criteria and none of the exclusion criteria, a researcher with randomization rights (only CD and DS) clicks the ‘Randomize’ button within 1 hour after informed consent was obtained. Randomization cannot be undone and is logged in the audit trail. The group allocation of patients is not visible for data-entry personnel. Due to the nature of the intervention, the study coordinators (CD, JV, DS, ME), dietitians and participants are not blinded to allocation.

Data will be collected at five time points; within 48 hours after admission (t ₀), at the 4 ^th day of admission (t ₁), at discharge (t ₂), one month after discharge to home (t ₃) and three months after discharge (t ₄). Table 2 provides an overview of the time, location, content of assessment and duration of data collection per time point. The measurements will be performed by a team of trained professionals according to the Dutch Medical Research Involving Subjects Act, ICH-GCP guidelines and Standard Operating Procedures. The team is supervised by the study coordinators (CD, JV, DS). Data will be collected on paper using Case Report Forms, which are stored at participating centres. Research team members enter the data in the online data management system Castor EDC (Castor EDC©, Amsterdam, The Netherlands). Personal information will be stored in password protected files, only accessible by the research team, according to the European General Data Protection Regulation.

In the development of the intensive dietetic care pathway, an Intervention Mapping (IM) approach was used. ^{15, 16} The framework consists of six steps and helped to process effective health behavioural theories and behavioural change techniques into the intervention. In step 1 of IM, we defined the problem ¹⁷ and performed an additional needs assessment with different disciplines involved in the treatment of older patients, to identify potential strategies ( publication in press). In step 2 and 3 of IM, a logical model of change was created to show which behavioural change techniques underlie parts of the intervention and their proposed change in the behavioural and health outcomes. Step 4 of IM consisted of the production of the intervention materials. For an extensive overview of the applications and their underlying behavioural changes techniques, see Table S1 and S2 in Supplementary file 1. ⁶⁹ In step 5 of IM, the implementation of the intervention will be performed. Actions needed for implementation are providing training to dietitians, physiotherapists, nurses, treating physicians, nutrition assistants and care givers around the patient. Online and offline training sessions will be created for this purpose. Step 6 of IM consists of the evaluation. Next to the evaluation of the effect of the intervention, also process and economic evaluation will be performed. To evaluate the implementation process, information on five process indicators will be assessed, being Recruitment/Reach, Dose received, Acceptability, Applicability, Implementation integrity. ^{18– 21} For the economic evaluation, information on quality of life, and the use and costs of healthcare will be assessed using validated questionnaires. ^{22, 23}

Control group

Participants in the control group will receive regular care and for the study only additional measurements will be performed (see Figure 1). Within current regular care, screening for risk of malnutrition is done with the SNAQ screening tool only. ²⁴ When the patient is found to be at risk, a dietitian is consulted by the nurse. When consulted, the dietitian specifies the nutritional status of the patient and develops a personalised protein and energy-enriched diet. The treatment goal is to reach the recommended protein and energy intake on the 4 ^th day of admission based on the Dutch guideline for malnutrition. ²⁵ Oral nutritional supplements or enteral nutrition via tube feeding can be used to achieve this goal. Follow-up of treatment differs greatly in practice. Some patients are lost to follow-up, some in-hospital dietitians perform a few telephone follow-ups, others hand over the patient to a primary care dietitian. If a patient has trouble with Activities of Daily Living (ADL) during hospitalisation, an in-hospital physiotherapist may be consulted. The focus of physiotherapy is usually ‘what must the patient be able to do, to be dischargeable’. A hand-over file is given to the patient at discharge, after which patients can decide for themselves if they want a follow-up treatment from a primary care physiotherapist.

Intervention group

For participants at risk of malnutrition in the intervention group, nutritional support will be initiated as soon as possible. Participants in this group will receive personalised intensive dietetic care (IDC) based on the scientific literature (see Table 3). ^{5, 25– 29} The nurses and nutritional assistant will be informed by the research team to start patients on a protein and energy-rich diet within 24 hours after admission. In addition, the dietitian will be informed by the study coordinators immediately so treatment can start within 48 hours. The treatment goal is to reach the recommended protein and energy intake on the 4 ^th day of admission and continuously after discharge. There is a set of smaller and larger adjustments that align consultation, interprofessional collaboration and communication. Determining the nutritional status, which is done at admission by a dietitian, are more specific than currently done in regular care (see Table 3). Oral nutritional support (ONS) or enteral nutrition via tube feeding can be used to achieve the protein and energy goal. The monitoring by the dietitian is intensified and supported by nutritional assistants, nurses, physiotherapists and informal caregivers. For all strategies applied within the intensive dietetic care, see Table 3. These strategies are based on the needs assessment and behavioural change theories and techniques (see Supplementary file 1, Table S1 and Table s2 ⁶⁹ ).

During hospitalisation

Hospital dietitian

The dietitian will have additional treatment time (90 minutes) for each patient to help achieve a tailored protein and energy rich diet. The extra time can be used to elaborate more on the importance of protein and mobility with the help of the knowledge-clip “Protein&Exercise” and infographic “Protein&Exercise”. During an interactive moment, the dietitian informs the family and their caregivers on the importance of protein intake and physical activity of the patient. ⁴³ Extra information on how they can support the patient in reaching their goals will be given. The dietitian encourages the patient to not lay in the hospital bed (with assistance of a nurse or physiotherapist if needed) during the day to stimulate physical activity. Other tools the dietitian may use during treatment are motivational interviewing, ONS or enteral tube feeding, focus on timing of protein around exercise and the “Protein&Exercise board” to create an overview of the daily goals. The extra treatment time can be used to collaborate the hospital physiotherapist to align the treatment goals; collaborating and communicating with the nurse and Nutrition Assistant (NA) about the participants’ food intake and treatment directly after first consult and every other day during hospitalisation can improve the participants’ adherence to the treatment plan. When the patient is discharged, the dietitian fills in the ProIntens medical file transfer and sends it to the dietitian in the primary care via secured e-mail. The dietitians will call each other to discuss details on the treatment of the patient within the first week post-discharge.

Hospital physiotherapist

During hospitalisation, the physiotherapist will be consulted at least once to assess which physical activity is safe and feasible for the patient, and to inform the patient on the results of the assessment. The physiotherapist will collaborate with the dietitian and aligns the treatment plans on timing of mobilisation and protein intake. When the patient will be discharged, the physiotherapist sends a medical file transfer to the primary care physiotherapist and/or dietitian.

Nutrition assistant (NA)

The nutrition assistant supports the patient in reaching the treatment goals by referring to the knowledge-clip and infographic “Protein&Exercise”. The NA encourages the patient to consume protein rich meals and products and provides protein rich snacks with a minimum of 5 g and optimally 10 g of protein immediately when requested by a patient. The NA knows the potential barriers to food intake and how to tackle them. During hospitalisation, the NA helps the patient monitoring their food intake with the Rate-a-Plate list or EiFit app (e-version of Rate-a-Plate). The NA provides feedback on nutritional status to the patient and the nurse daily and has direct contact with the dietitian every other day. If needed, the NA helps the patient with adjusting the “Protein&Exercise” board.

After discharge

Primary care dietitian

The dietitian in primary care receives the medical file from the hospital dietitian via secured e-mail and will contact the hospital dietitian to receive additional information on the patient within the first week after discharge. Within the same first few days, the dietitian will contact the patient to schedule the first home visit within the first week after discharge. In addition, the dietitian actively refers the patient to a physiotherapist to continue the support on mobilisation. ^{38– 40} During the first visit, the follow-up consultations will be planned (weekly contact for four weeks, thereafter biweekly until three months after discharge; see Table 4). During the first home visit, the dietitian may adjust the treatment plan to suit the participant’s home setting. When needed, they can refer to knowledge-clip and infographic “Protein&Exercise” to repeat the importance to continue the treatment plan. During the three months of treatment, the dietitian aids patient learning to monitor food intake based on Rate-a-Plate list or EiFit app and evaluate on intake during every visit. At every visit, the dietitian mentions the importance of physical activity to the participants, family and informal caregivers and encourages the participant to be physically active. The dietitian may refer to the exercise program in the folder. When needed, the dietitian can inform the participant about home delivery and other social services, which are available in the discharge folder.

Primary care physiotherapist (if consulted)

The consultation of the physiotherapist is dependent on insurance of the patient. When consulted, the physiotherapist receives information on the participant’s participation in the study, and receives the patient files via secured e-mail. During the follow-up the dietitian and physiotherapist collaborate to align the interprofessional treatment plans on timing of mobilisation and protein intake.

Community nurse (if consulted)

When a community nurse is consulted in support of recovery of the patient at home, they will receive information on the presence of malnutrition and the treatment plans, which are available in the discharge folder of the patient. The dietitian and physiotherapist will be notified of the involvement of the community nurse. The community nurse has knowledge on protein rich products and encourages the patient to consume protein rich products. Also, the community nurse can encourage the patient to be active. When the community nurse suspects a low intake, based on the Rate-a-Plate list or the EiFit application, they will notify the dietitian.

Figure 1 provides a detailed overview of the primary and secondary outcomes measured at each time point.

Primary outcome

The primary outcome is the between-groups difference in the mean change on the Short Physical Performance Battery from admission (t ₁) to three months post-discharge (t ₄). ⁴⁴ This battery consists of the following tasks: balance stance in three different positions: a side-by-side, semi tandem, and full tandem, walking three or four meters to measure gait speed, and rising from a chair and return, five times as quickly as possible. The total scores can range from 0 to 12 (best performance). If a patient is unable to perform a test, 0 points will be noted. A change of 1 point on the total score has been demonstrated to be clinically meaningful. ^{45, 46}

Secondary outcomes 1)

Physical functioning and performance

Handgrip strength Handgrip strength is an objective measure of muscle strength in the upper body. Three consecutive measures of handgrip strength (kg) of both hands will be recorded to the nearest 0.5 kg using a hand dynamometer (Jamar). Standardized vocal stimulation will be given, according to the Standard Operating Procedure (SOP) of the Nutritional Assessment Platform (NAP). ⁴⁷ The highest score will be used as a parameter for muscle strength.

Knee extension strength The knee extension strength is measured with a hand-held dynamometer (MicroFET2). The strength is measured in kilograms (kg). Participants will be asked to be seated on an elevated chair with their knees at a 90° angle, feet off the floor and hands on their thighs. The dynamometer will be placed proximal to the talus and malleolus (ankle). The participants will be asked to slowly increase their muscle force to a maximum effort which will need to be sustained for three seconds. The procedure will be repeated two times for each leg. The highest score will be used as a parameter for lower leg strength. ⁴⁸

Physical activity Daily physical activity will be registered using a three-dimensional Physical Activity Monitor (PAM). The PAM is a compact, battery-operated electronic accelerometer. Participants are requested to wear the accelerometer around their ankle during hospital stay and a period of three days before t ₃ and t ₄. The PAM registers the amount of movement of the user, converted into number of steps and the Pam Score. The Pam Score is the ratio between activity induced energy expenditure and resting metabolic rate.

Activities of Daily Living (ADL) To assess ADL, the Barthel Index will be used. The questionnaire consists of 10 items on an ordinal scale that relate to the preceding 24-48 hours. A high score on the Barthel Index corresponds to a high degree of independence. ⁴⁹

Pain and fatigue A Numeric Rating Scale (NRS) will be used to assess pain and fatigue. The NRS for pain is a validated continuous scale with a score range between zero and ten, where zero represents no pain and ten the worst pain possible. ^{50, 51} The NRS for fatigue is a continuous scale with a score range between zero and ten, where zero represents no fatigue and ten the worst fatigue possible. ⁵²

Sleep quality To assess sleep quality two questions of the Pittsburgh Sleep Quality Index (PSQI), a self-reported questionnaire, are used, being ‘ During the past month, how would you rate your sleep quality overall?’ and ‘ During the past month, how often have you taken medicine (prescribed or “over the counter”) to help you sleep?’. Both questions have a range of 0 (better) to 4 (worse). ⁵³

To assess whether the implementation of the intervention was successful, a process evaluation will be conducted. A total of five process indicators are identified, being Recruitment/Reach, Dose received, Acceptability, Applicability, Implementation integrity. ^{18– 21} Recruitment/reach describes the procedures used to approach and attract participants and the proportion of intended target audience that participated in the intervention. This latter information will be obtained from the in- and exclusion flowchart. Dose received describes to which extent the participants received the intervention, i.e. number of consultations by dietitian and physiotherapist, discussed items during consultation, times intervention items are used. This information will be obtained from the cost questionnaire and surveys send to dietitian. Acceptability describes the degree of satisfaction of participants and healthcare professionals with the intervention. This information will be obtained through semi-structured interviews in a random subsample of participants who completed the study and healthcare professionals who carried out the intervention. Applicability describes the feasibility of implementing the intervention from the healthcare professional’s point of view. This information will be obtained through semi-structured interviews. Implementation integrity describes the degree to which the intervention was implemented as intended. This information will be obtained through semi-structured interviews with healthcare professionals. In addition, structured observations will be conducted by the research staff and registration forms of the health care professionals will be checked.

Summary statistics (mean, median, standard deviation, interquartile range and frequency distribution) will be generated for baseline characteristics. Descriptive data will be used to assess time trends.

SPPB total scores will be visualised over time using mean and standard error of the mean per treatment group and period or with median and interquartile range if needed. Differences between treatment groups in the primary outcome parameter will be analysed using generalized linear mixed models. A dummy variable for time will be created. Important prognostic baseline values will be used as covariates. Secondary outcome measures will be visualised over time using mean and standard error of the mean per treatment group or with median and interquartile range if needed. Changes within and between groups will be modelled with linear mixed models. All data will be analysed according to the intention-to-treat principle. Missing data will be handled with linear mixed models. An α-level of 0.05 will determine statistical significance. Data validation is performed by blinded project leaders (HK, MT). Further analyses will be stated in the Statistical Analysis Plan which will be finalised and put in the public domain via Figshare repository before unblinding of the dataset.

The economic evaluation will be done from a societal perspective according to the intention-to-treat principle. Following the guidelines for economic evaluations, provided by the Dutch Healthcare Institute, standard prices are used to calculate costs. ⁶⁴ Both a cost-effectiveness analysis with SPPB as the outcome measure and a cost-utility analysis with the EQ-5D5L and ASCOT as outcome measures will be performed. Missing cost and effect data will be imputed using multiple imputation with predictive mean matching. Cost and effect differences will be estimated using linear mixed models in combination with bootstrapping. Statistical uncertainty surrounding the incremental cost-effectiveness ratio will be estimated using bootstrapping. Cost-effectiveness acceptability curves will be estimated to show the probability that the intervention is cost-effective in comparison with usual care for different willingness-to-pay thresholds (i.e., the maximum amount of money society is willing to pay per unit of effect extra).

This study is approved by the Medical Ethics Committee (METc) of the Amsterdam University Medical Center (AUMC), location VUmc in The Netherlands (Protocol ID: 2019.689 (approved on 10 ^th of July 2020) and Amendment ID: A2020.361 (approved on 26 ^th of July 2021)). Written informed consent is obtained from all participants before any data is collected. Insurance has been taken out for compensation to those who suffer harm form the trial participation. The research is performed according to the Dutch Medical Research Involving Human Subjects Act (WMO) and the principles of the Declaration of Helsinki (64 ^th W MA General Assembly, Fortaleza, Brazil, October 2013).

Not applicable.

Recruitment of participants will continue until 31-12-2022. After completion of the trial, data will be stored and shared through Figshare.

The study has finished recruiting subjects on the 31 ^st of December 2022.